具体实施方式

下面将参照附图更详细地描述本公开的实施例。虽然附图中显示了本公开的某些实施例，然而应当理解的是，本公开可以通过各种形式来实现，而且不应该被解释为限于这里阐述的实施例。相反，提供这些实施例是为了更加透彻和完整地理解本公开。应当理解的是，本公开的附图及实施例仅用于示例性作用，并非用于限制本公开的保护范围。

另外还需要说明的是，为了便于描述，附图中仅示出了与有关发明相关的部分。在不冲突的情况下，本公开中的实施例及实施例中的特征可以相互组合。

需要注意，本公开中提及的“第一”、“第二”等概念仅用于对不同的装置、模块或单元进行区分，并非用于限定这些装置、模块或单元所执行的功能的顺序或者相互依存关系。

需要注意，本公开中提及的“一个”、“多个”的修饰是示意性而非限制性的，本领域技术人员应当理解，除非在上下文另有明确指出，否则应该理解为“一个或多个”。

本公开实施方式中的多个装置之间所交互的消息或者信息的名称仅用于说明性的目的，而并不是用于对这些消息或信息的范围进行限制。

下面将参考附图并结合实施例来详细说明本公开。

图1示出了根据本公开的图像清洗方法的一些实施例的流程100。该图像清洗方法，包括以下步骤：

步骤101，控制采集车，以针对限速牌进行数据采集，得到限速牌图像序列。

在一些实施例中，上述图像清洗方法的执行主体（例如，电子设备）可以控制采集车，以针对限速牌进行数据采集，得到限速牌图像序列。其中，上述限度牌图像数据序列可以是按照上述采集车对限速牌进行采集时落盘的时间顺序进行排序的图像序列。落盘可以是将采集到的限速牌图像序列存储至磁盘。落盘的时间可以是上述限速牌图像存储至磁盘中的时间。上述采集车可以是对上述限速牌进行采集的车辆。例如，上述采集车可以是无人驾驶车辆。

在一些实施例的一些可选的实现方式中，上述控制采集车，以针对限速牌进行数据采集，得到限速牌图像序列，可以包括以下步骤：

第一步，控制上述采集车，以针对限速牌进行数据采集，得到采集图像序列。其中，上述采集图像序列可以是对限速牌进行采集得到的图像序列。上述采集图像序列的顺序可以是按照落盘的时间顺序进行排序得到序列。

第二步，对上述采集图像序列进行通道划分，得到多个通道图像序列。其中，上述多个通道图像序列中的通道图像可以是仅包括一种颜色通道的图像。例如，多个通道图像序列可以包括：红色通道图像序列、绿色通道图像序列和蓝色通道图像序列。

第三步，对上述多个通道图像序列中的每个通道图像，执行以下确定步骤：

子步骤1，对上述通道图像进行平滑处理，得到平滑图像。其中，上述平滑处理可以是去除上述通道图像中高频像素的图像。实践中，上述执行主体可以首先，利用傅立叶变换，将上述通道图像转换到频率域，得到频域图像。其中，上述频域图像可以是将位于空间域的通道图像进行傅立叶变换至频率域得到的图像。上述频率域可以表征通道图像灰度的变化情况。上述空间域可以是二维平面坐标系。然后，利用低通滤波器，对上述频域图像进行去噪处理，得到平滑图像序列。

子步骤2，对上述平滑图像进行频域转换，得到空间域图像。其中，上述空间域图像可以是图像中的像素位于二维平面坐标系中的图像。实践中，上述执行主体可以利用反傅里叶变换，对上述平滑图像进行频域转换，得到空间域图像。

子步骤3，确定上述空间域图像的反射率。其中，上述反射率可以是物体本身对光进行的反射的能量和投射到物体上的光的总能量的比值。如图2所示，空间域图像是由入射光和反射率的乘积组成的图像。I（x，y）表示空间域图像。L（x，y）表示入射光。 R（x，y）表示反射率。上述入射光照可以是对限速牌进行拍摄时外部光照。实践中，上述执行主体可以首先，对上述空间域图像中的像素值进行归一化处理，得到归一化数值。然后，对上述归一化数值进行高斯卷积滤波处理，得到上述空间域图像的反射率。

第四步，对所得到的多个反射率序列进行通道融合，得到融合图像序列。其中，上述融合图像序列可以是融合三种颜色通道对应反射率得到的限速牌图像序列。

第五步，对上述融合图像序列进行均衡化处理，得到限速牌图像序列。实践中，上述执行主体可以利用直方图均衡法，对上述融合图像序列进行均衡化处理，得到限速牌图像序列。

上述技术方案及其相关内容作为本公开的实施例的一个发明点，解决了背景技术提及的技术问题二“仅时空域图像滤波器或者频率域图像滤波器，对恶劣天气下的采集的图像数据进行滤波处理，导致图像中物体的边缘信息不清晰，进而导致自动驾驶车辆识别的准确度和安全性较低”。导致自动驾驶车辆识别的准确度和安全性较低的因素往往如下：仅时空域图像滤波器或者频率域图像滤波器，对恶劣天气下的采集的图像数据进行滤波处理，导致图像中物体的边缘信息不清晰，进而导致自动驾驶车辆识别的准确度和安全性较低。如果解决了上述因素，就能达到提高图像中物体的边缘信息的清晰度和减少图像整体亮度差异的效果。为了达到这一效果，本公开首先，控制上述采集车，以针对限速牌进行数据采集，得到采集图像序列。其次，对上述采集图像序列进行通道划分，得到多个通道图像序列。在这里，进行通道划分可以在细粒度程度上对图像进行增强处理，可以得到更加清晰、无噪声的图像。随后，对上述通道图像进行平滑处理，得到平滑图像。在这里，平滑处理有利于减弱图像中的噪声，更加突出图像中的重要信息。对上述平滑图像进行频域转换，得到空间域图像。在这里，由频率域转换到空间域更加有利于确定后续图像的反射率。确定上述空间域图像的反射率。然后，对所得到的多个反射率序列进行通道融合，得到融合图像序列。在这里，确定反射率可以增强去除恶劣天气中去雾效果，提高图像的清晰度和边缘化信息。最后，对上述融合图像序列进行均衡化处理，得到限速牌图像序列。在这里，均衡化处理可以增加图像的对比度和清晰度，以及减少图像中的噪声。由此，上述技术方案结合频率域和空间域去噪方法，可以实现提高恶劣天气下采集的图像序列中图像的清晰度和物体的边缘信息，进而提高自动驾驶车辆识别的准确度和安全性。

步骤102，从限速牌图像序列中选取样本限速牌图像序列，以及执行以下确定步骤：

步骤1021，将样本限速牌图像序列依次输入至初始目标检测模型，得到概率数值序列。

在一些实施例中，上述执行主体可以将上述样本限速牌图像序列依次输入至初始目标检测模型，得到概率数值序列。其中，上述概率数值序列中的概率数值表征样本限速牌图像中存在上述限速牌的概率数值。上述样本限速牌图像序列可以是用于对上述初始目标检测模型进行训练的图像序列。上述初始目标检测模型可以包括但不限于：YOLO（You OnlyLook Once）模型，SSD（Single Shot MultiBox Detector）模型。

步骤1022，根据概率数值序列，确定目标时间点。

在一些实施例中，上述执行主体可以根据上述概率数值序列，确定目标时间点。其中，上述目标时间点可以是上述概率数值序列中大于等于预设概率阈值，且位于初始位置的概率数值对应的落盘时间点。初始位置可以是第一个位置。

作为示例，上述执行主体可以首先，将上述概率数值序列中每个概率数值与预设概率阈值进行对比，以生成对比结果，得到对比结果序列。其次，确定上述对比结果序列中大于等于预设概率阈值的对比结果，作为目标结果序列。最后，将位于上述目标结果序列中初始位置的对比结果对应的概率数值对应的采集时间点，确定为目标时间点。

在一些实施例的一些可选的实现方式中，上述根据上述概率数值序列，确定目标时间点，可以包括以下步骤：

第一步，从上述概率数值序列中筛选出大于等于上述预设概率阈值的概率数值，得到目标概率数值序列。其中，预设概率阈值可以是0.75。

第二步，将上述目标概率数值序列对应的样本限速牌图像序列，确定为目标样本限速牌图像序列。

第三步，将上述目标样本限速牌图像序列中位于初始位置的目标样本限速牌图像对应的采集时间点，确定为目标时间点。

步骤1023，根据目标时间点，创建初始可变窗口。

在一些实施例中，上述执行主体可以根据上述目标时间点，创建初始可变窗口。其中，上述初始可变窗口用于确定待清洗的样本限速牌图像的窗口。

作为示例，上述执行主体可以针对上述目标时间点对应的样本限速牌图像创建初始可变窗口。

在一些实施例的一些可选的实现方式中，上述根据上述目标时间点，创建初始可变窗口，可以包括以下步骤：

第一步，将与上述目标时间点之间对应差值位于第一预设时间范围内的时间点，确定为起始时间点。例如，上述第一预设时间范围内可以是1秒至3秒。与上述目标时间点之间对应差值可以是3秒。

第二步，从上述样本限速牌图像序列中选取出与上述起始时间点相同的采集时间点对应的样本限速牌图像，作为起始样本限速牌图像。其中，上述采集时间点可以是采集车对限速牌进行采集、落盘至内存中的时间点。

第三步，将与上述目标时间点之间对应差值位于第二预设时间范围内的时间点，确定为终止时间点。例如，上述第二预设时间范围可以是1秒至2秒。与上述时间点之间对应差值可以是1秒。

第四步，从上述样本限速牌图像序列中选取出与上述终止时间点相同的采集时间点对应的样本限速牌图像，作为终止样本限速牌图像。

第五步，将位于上述样本限速牌图像序列中的、从上述起始样本限速牌图像至上述终止样本限速牌图像之间的样本限速牌图像，确定为上述初始可变窗口包括的样本限速牌图像序列。

步骤1024，根据初始可变窗口，生成至少一个图像序列。

在一些实施例中，上述执行主体可以根据上述初始可变窗口，生成至少一个图像序列。其中，上述至少一个图像序列可以是至少一个初始可变窗口包括的样本限速牌图像序列。

作为示例，上述执行主体可以针对上述样本限速牌图像序列对应的概率数值中大于等于预设概率阈值的概率数值创建初始可变窗口，得到的至少一个图像序列。

在一些实施例的一些可选的实现方式中，上述根据上述初始可变窗口，生成至少一个图像序列，可以包括以下步骤：

第一步，基于第一概率数值序列，执行以下序列确定步骤：

子步骤1，响应于确定上述初始可变窗口不包括与上述序列终止时间点相同的采集时间点对应的样本限速牌图像序列，对上述初始可变窗口进行扩展，得到扩展可变窗口和扩展窗口。其中，上述扩展窗口可以是上述扩展可变窗口大于上述初始可变窗口的部分对应的窗口。上述扩展可变窗口可以是对上述初始可变窗口的终止位置对应的时间点进行扩展得到的窗口。上述第一概率数值序列是上述概率数值序列、从上述目标时间点至序列终止时间点之间的概率数值序列，上述序列终止时间点可以是位于上述概率数值序列终止位置的概率数值对应的时间点。

子步骤2，响应于确定上述扩展窗口包括的样本限速牌图像序列对应的概率数值序列中不存在大于等于上述预设概率阈值的概率数值，将上述扩展可变窗口包括的样本限速牌图像序列添加至至少一个预设图像序列。其中，上述至少一个预设图像序列可以是预先设定的、包括至少一个图像序列的图像序列。

子步骤3，响应于确定第二概率数值序列中存在大于等于上述预设概率阈值的第二概率数值，从上述第二概率数值序列中筛选出大于等于上述预设概率数值的第二概率数值，得到目标筛选概率数值序列。其中，上述第二概率数值序列是位于上述概率数值序列、从扩展概率数值至上述序列终止时间点对应的概率数值之间的概率数值序列。上述扩展概率数值是所得到的扩展可变窗口中位于终止位置的样本限速牌图像序列对应的概率数值。

子步骤4，针对上述目标筛选概率数值序列中位于起始位置的概率数值对应的样本限速牌图像，创建样本可变窗口。其中，上述样本可变窗口可以用于确定待清洗的样本限速牌图像的窗口。

子步骤5，响应于确定上述样本可变窗口包括与上述序列终止时间点相同的采集时间点对应的样本限速牌图像，将上述至少一个预设图像序列，确定为上述至少一个图像序列。

第二步，响应于确定样本可变窗口不包括与上述序列终止时间点相同的采集时间点对应的样本限速牌图像，将位于上述概率数值序列、从窗口终止时间点至上述序列终止时间点之间的概率数值序列，确定为第一概率数值序列，以及再次执行上述序列确定步骤。其中，上述窗口终止时间点是位于样本可变窗口终止位置的样本限速牌图像对应的时间点。

在一些实施例的一些可选的实现方式中，上述对上述初始可变窗口进行扩展，得到扩展可变窗口和扩展窗口，包括：

第一步，基于上述初始可变窗口，执行以下扩展步骤：

子步骤1，将上述目标时间点至上述终止时间点之间包括的样本限速牌图像对应的概率数值，确定为窗口概率数值序列。其中，上述窗口概率数值序列可以是目标时间点和终止时间点之间包括的样本限速牌图像序列对应的概率数值。

子步骤2，响应于确定上述窗口概率数值序列中存在大于等于上述预设概率阈值的窗口概率数值，确定大于等于上述预设概率阈值的至少一个窗口概率数值，作为目标窗口概率数值序列。其中，上述目标窗口概率数值序列可以是位于上述窗口概率数值序列中的、大于等于上述预设概率阈值的概率数值序列。

子步骤3，确定位于上述目标窗口概率数值序列终止位置的目标窗口概率数值，作为目标终止窗口概率数值。

子步骤4，将上述目标终止窗口概率数值的采集时间点，确定为调节时间点。

子步骤5，根据上述调节时间点，对上述初始可变窗口进行扩展，得到扩展可变窗口和上述扩展窗口。其中，上述扩展可变窗口可以是对初始可变窗口的终止位置对应的时间点向后扩展预设时间的窗口。例如，上述预设时间可以是1秒。

作为示例，上述执行主体从上述调节时间点为起始时间点，对上述初始可变窗口的终止位置对应的时间点向后扩展1秒，得到扩展可变窗口。

子步骤6，响应于确定上述扩展窗口包括的样本限速牌图像序列对应概率数值序列中不存在大于等于上述预设概率阈值的概率数值，结束上述扩展步骤。

可选地，上述执行主体还可以执行以下步骤：

响应于确定扩展窗口包括的样本限速牌图像序列对应的概率数值序列中存在大于等于上述预设概率阈值的概率数值，将初始可变窗口中位于终止位置的样本限速牌图像的采集时间点确定为目标时间点，将扩展可变窗口确定为初始可变窗口，以及再次执行上述扩展步骤。

可选地，在所述响应于确定上述初始可变窗口不包括与上述序列终止时间点相同的采集时间点对应的样本限速牌图像序列，对上述初始可变窗口进行扩展，得到扩展可变窗口和扩展窗口之后，所述方法还包括：

响应于确定上述初始可变窗口包括与上述序列终止时间点相同的采集时间点对应的样本限速牌图像序列，将上述初始可变窗口包括的样本限速牌图像序列添加至上述至少一个预设图像序列。

步骤1025，对至少一个图像序列进行清洗处理，得到处理后限速牌图像序列。

在一些实施例中，上述执行主体可以对上述至少一个图像序列进行清洗处理，得到处理后限速牌图像序列。其中，上述处理后限速牌图像序列可以是概率阈值小于预设概率阈值、但是图像中包括限速牌的图像序列。上述清洗处理可以包括但不限于：利用统计法对至少一个图像序列进行清洗处理、人工检测清洗处理。

步骤1026，确定概率数值序列中大于等于预设概率阈值的概率数值的数目与概率数值序列中概率数值的数目的比值。

在一些实施例中，上述执行主体可以确定上述概率数值序列中大于等于预设概率阈值的概率数值的数目与上述概率数值序列中概率数值的数目的比值。其中，上述预设概率阈值可以表征初始目标检测模型的训练完成。例如，上述预设概率阈值可以是0.75。

步骤1027，响应于确定比值大于等于预设阈值，将初始可变窗口确定为目标可变窗口。

在一些实施例中，上述执行主体可以响应于确定上述比值大于等于预设阈值，将上述初始可变窗口确定为目标可变窗口。其中，上述目标可变窗口可以是与初始目标检测模型相对应的窗口。上述目标可变窗口可以是随着初始目标模型的准确度的提高而可变窗口的长度减小的窗口。

步骤103，响应于确定比值小于预设阈值，对初始可变窗口和初始目标检测模型进行调整，将处理后限速牌图像序列作为样本限速牌图像序列，分别将调整后的初始可变窗口和调整后的初始目标检测模型确定为初始可变窗口和初始目标检测模型，以及再次执行确定步骤。

在一些实施例中，上述执行主体可以响应于确定比值小于上述预设阈值，对初始可变窗口和初始目标检测模型进行调整，将上述处理后限速牌图像序列作为样本限速牌图像序列，分别将调整后的初始可变窗口和调整后的初始目标检测模型确定为初始可变窗口和初始目标检测模型，以及再次执行确定步骤。

本公开的上述各个实施例中具有如下有益效果：本公开的一些实施例的图像清洗方法利用可调节大小的可变窗口和目标检测模型可以降低图像清洗的漏检率、提高图像清洗的效率。具体来说，造成相关的自动驾驶的安全性降低的原因在于：人工清洗的效率较低，并且清洗周期和清洗成本较高，而利用预先训练好的目标检测模型对限速牌检测的准确率较低，并且对于一些边缘场景的图像的漏检率较高，导致图像清洗效率较低。基于此，本公开的一些实施例的图像清洗方法可以首先，控制采集车，以针对限速牌进行数据采集，得到限速牌图像序列。在这里，限速牌图像序列便于结合初始可变窗口降低漏检率，以及便于后续对可变窗口和目标检测模型进行调节。然后，从上述限速牌图像序列中选取样本限速牌图像序列，以及执行以下确定步骤：将上述样本限速牌图像序列依次输入至初始目标检测模型，得到概率数值序列，其中，上述概率数值序列中的概率数值表征样本限速牌图像中存在上述限速牌的概率数值。在这里，得到的概率数值序列便于后续确定初始目标检测模型的准确率。根据上述概率数值序列，确定目标时间点。在这里，得到目标时间点便于创建初始可变窗口，以及降低样本限速牌图像序列清洗的漏检率，得到限速牌的边缘图像数据有利于提高模型检测率。根据上述目标时间点，创建初始可变窗口，其中，上述初始可变窗口用于确定待清洗的样本限速牌图像的窗口。在这里，创建初始可变窗口可以获取样本限速牌图像序列中的边缘图像，提高样本限速牌图像序列清洗的漏检率。根据上述初始可变窗口，生成至少一个图像序列。在这里，得到至少一个图像序列便于后续进行清洗处理，得到边缘图像。随后，对上述至少一个图像序列进行清洗处理，得到处理后限速牌图像序列。在这里，对至少一个图像序列进行清洗可以提高目标检测模型的检测准确度，减少样本限速牌图像序列的清洗周期和清洗成本。确定上述概率数值序列中大于等于预设概率阈值的概率数值的数目与上述概率数值序列中概率数值的数目的比值；响应于确定上述比值大于等于预设阈值，将上述初始可变窗口确定为目标可变窗口。最后，响应于确定比值小于上述预设阈值，对初始可变窗口和初始目标检测模型进行调整，将上述处理后限速牌图像序列作为样本限速牌图像序列，分别将调整后的初始可变窗口和调整后的初始目标检测模型确定为初始可变窗口和初始目标检测模型，以及再次执行确定步骤。在这里，基于目标检测模型和可变窗口可以降低样本限速牌图像序列清洗的漏检率，对初始可变窗口和初始目标检测模型进行调节可以提高模型和图像清洗的效率，降低清洗周期和清洗成本。由此可得，该图像清洗方法利用可调节大小的可变窗口和目标检测模型可以降低图像清洗的漏检率、提高图像清洗的效率和目标检测模型的检测效率。

进一步参考图3，作为对上述各图所示方法的实现，本公开提供了一种图像清洗装置的一些实施例，这些装置实施例与图1所示的那些方法实施例相对应，该图像清洗装置具体可以应用于各种电子设备中。

如图3所示，一种图像清洗装置300包括：采集单元301、执行单元302和调整单元303。其中，采集单元301被配置成：控制采集车，以针对限速牌进行数据采集，得到限速牌图像序列。执行单元302被配置成：从上述限速牌图像序列中选取样本限速牌图像序列，以及执行以下确定步骤：将上述样本限速牌图像序列依次输入至初始目标检测模型，得到概率数值序列，其中，上述概率数值序列中的概率数值表征样本限速牌图像中存在上述限速牌的概率数值；根据上述概率数值序列，确定目标时间点；根据上述目标时间点，创建初始可变窗口，其中，上述初始可变窗口用于确定待清洗的样本限速牌图像的窗口；根据上述初始可变窗口，生成至少一个图像序列；对上述至少一个图像序列进行清洗处理，得到处理后限速牌图像序列；确定上述概率数值序列中大于等于预设概率阈值的概率数值的数目与上述概率数值序列中概率数值的数目的比值；响应于确定上述比值大于等于预设阈值，将上述初始可变窗口确定为目标可变窗口。调整单元303被配置成：响应于确定比值小于上述预设阈值，对初始可变窗口和初始目标检测模型进行调整，将上述处理后限速牌图像序列作为样本限速牌图像序列，分别将调整后的初始可变窗口和调整后的初始目标检测模型确定为初始可变窗口和初始目标检测模型，以及再次执行确定步骤。

可以理解的是，图像清洗装置300中记载的诸单元与参考图1描述的方法中的各个步骤相对应。由此，上文针对方法描述的操作、特征以及产生的有益效果同样适用于图像清洗装置300及其中包含的单元，在此不再赘述。

下面参考图4，其示出了适于用来实现本公开的一些实施例的电子设备（例如，电子设备）400的结构示意图。图4示出的电子设备仅仅是一个示例，不应对本公开的实施例的功能和使用范围带来任何限制。

如图4所示，电子设备400可以包括处理装置（例如中央处理器、图形处理器等）401，其可以根据存储在只读存储器（ROM）402中的程序或者从存储装置408加载到随机访问存储器（RAM）403中的程序而执行各种适当的动作和处理。在RAM403中，还存储有电子设备400操作所需的各种程序和数据。处理装置401、ROM 402以及RAM 403通过总线404彼此相连。输入/输出（I/O）接口305也连接至总线404。

通常，以下装置可以连接至I/O接口405：包括例如触摸屏、触摸板、键盘、鼠标、摄像头、麦克风、加速度计、陀螺仪等的输入装置406；包括例如液晶显示器（LCD）、扬声器、振动器等的输出装置407；包括例如磁带、硬盘等的存储装置408；以及通信装置409。通信装置409可以允许电子设备400与其他设备进行无线或有线通信以交换数据。虽然图4示出了具有各种装置的电子设备400，但是应理解的是，并不要求实施或具备所有示出的装置。可以替代地实施或具备更多或更少的装置。图4中示出的每个方框可以代表一个装置，也可以根据需要代表多个装置。

特别地，根据本公开的一些实施例，上文参考流程图描述的过程可以被实现为计算机软件程序。例如，本公开的一些实施例包括一种计算机程序产品，其包括承载在计算机可读介质上的计算机程序，该计算机程序包含用于执行流程图所示的方法的程序代码。在这样的一些实施例中，该计算机程序可以通过通信装置409从网络上被下载和安装，或者从存储装置408被安装，或者从ROM 402被安装。在该计算机程序被处理装置401执行时，执行本公开的一些实施例的方法中限定的上述功能。

需要说明的是，本公开的一些实施例上述的计算机可读介质可以是计算机可读信号介质或者计算机可读存储介质或者是上述两者的任意组合。计算机可读存储介质例如可以是——但不限于——电、磁、光、电磁、红外线、或半导体的系统、装置或器件，或者任意以上的组合。计算机可读存储介质的更具体的例子可以包括但不限于：具有一个或多个导线的电连接、便携式计算机磁盘、硬盘、随机访问存储器（RAM）、只读存储器（ROM）、可擦式可编程只读存储器（EPROM或闪存）、光纤、便携式紧凑磁盘只读存储器（CD-ROM）、光存储器件、磁存储器件、或者上述的任意合适的组合。在本公开的一些实施例中，计算机可读存储介质可以是任何包含或存储程序的有形介质，该程序可以被指令执行系统、装置或者器件使用或者与其结合使用。而在本公开的一些实施例中，计算机可读信号介质可以包括在基带中或者作为载波一部分传播的数据信号，其中承载了计算机可读的程序代码。这种传播的数据信号可以采用多种形式，包括但不限于电磁信号、光信号或上述的任意合适的组合。计算机可读信号介质还可以是计算机可读存储介质以外的任何计算机可读介质，该计算机可读信号介质可以发送、传播或者传输用于由指令执行系统、装置或者器件使用或者与其结合使用的程序。计算机可读介质上包含的程序代码可以用任何适当的介质传输，包括但不限于：电线、光缆、RF（射频）等等，或者上述的任意合适的组合。

在一些实施方式中，客户端、服务器可以利用诸如HTTP（Hyper Text TransferProtocol，超文本传输协议）之类的任何当前已知或未来研发的网络协议进行通信，并且可以与任意形式或介质的数字数据通信（例如，通信网络）互连。通信网络的示例包括局域网（“LAN”），广域网（“WAN”），网际网（例如，互联网）以及端对端网络（例如，adhoc端对端网络），以及任何当前已知或未来研发的网络。

上述计算机可读介质可以是上述电子设备中所包含的；也可以是单独存在，而未装配入该电子设备中。上述计算机可读介质承载有一个或者多个程序，当上述一个或者多个程序被该电子设备执行时，使得该电子设备：控制采集车，以针对限速牌进行数据采集，得到限速牌图像序列；从上述限速牌图像序列中选取样本限速牌图像序列，以及执行以下确定步骤：将上述样本限速牌图像序列依次输入至初始目标检测模型，得到概率数值序列，其中，上述概率数值序列中的概率数值表征样本限速牌图像中存在上述限速牌的概率数值；根据上述概率数值序列，确定目标时间点；根据上述目标时间点，创建初始可变窗口，其中，上述初始可变窗口用于确定待清洗的样本限速牌图像的窗口；根据上述初始可变窗口，生成至少一个图像序列；对上述至少一个图像序列进行清洗处理，得到处理后限速牌图像序列；确定上述概率数值序列中大于等于预设概率阈值的概率数值的数目与上述概率数值序列中概率数值的数目的比值；响应于确定上述比值大于等于预设阈值，将上述初始可变窗口确定为目标可变窗口；响应于确定比值小于上述预设阈值，对初始可变窗口和初始目标检测模型进行调整，将上述处理后限速牌图像序列作为样本限速牌图像序列，分别将调整后的初始可变窗口和调整后的初始目标检测模型确定为初始可变窗口和初始目标检测模型，以及再次执行确定步骤。

可以以一种或多种程序设计语言或其组合来编写用于执行本公开的一些实施例的操作的计算机程序代码，上述程序设计语言包括面向对象的程序设计语言—诸如Java、Smalltalk、C++，还包括常规的过程式程序设计语言—诸如“C”语言或类似的程序设计语言。程序代码可以完全地在用户计算机上执行、部分地在用户计算机上执行、作为一个独立的软件包执行、部分在用户计算机上部分在远程计算机上执行、或者完全在远程计算机或服务器上执行。在涉及远程计算机的情形中，远程计算机可以通过任意种类的网络——包括局域网（LAN）或广域网（WAN）——连接到用户计算机，或者，可以连接到外部计算机（例如利用因特网服务提供商来通过因特网连接）。

附图中的流程图和框图，图示了按照本公开各种实施例的系统、方法和计算机程序产品的可能实现的体系架构、功能和操作。在这点上，流程图或框图中的每个方框可以代表一个模块、程序段、或代码的一部分，该模块、程序段、或代码的一部分包含一个或多个用于实现规定的逻辑功能的可执行指令。也应当注意，在有些作为替换的实现中，方框中所标注的功能也可以以不同于附图中所标注的顺序发生。例如，两个接连地表示的方框实际上可以基本并行地执行，它们有时也可以按相反的顺序执行，这依所涉及的功能而定。也要注意的是，框图和/或流程图中的每个方框、以及框图和/或流程图中的方框的组合，可以用执行规定的功能或操作的专用的基于硬件的系统来实现，或者可以用专用硬件与计算机指令的组合来实现。

描述于本公开的一些实施例中的单元可以通过软件的方式实现，也可以通过硬件的方式来实现。所描述的单元也可以设置在处理器中，例如，可以描述为：一种处理器包括采集单元、执行单元和调整单元。其中，这些单元的名称在某种情况下并不构成对该单元本身的限定，例如，采集单元还可以被描述为“控制采集车，以针对限速牌进行数据采集，得到限速牌图像序列的单元”。

本文中以上描述的功能可以至少部分地由一个或多个硬件逻辑部件来执行。例如，非限制性地，可以使用的示范类型的硬件逻辑部件包括：现场可编程门阵列（FPGA）、专用集成电路（ASIC）、专用标准产品（ASSP）、片上系统（SOC）、复杂可编程逻辑设备（CPLD）等等。

以上描述仅为本公开的一些较佳实施例以及对所运用技术原理的说明。本领域技术人员应当理解，本公开的实施例中所涉及的发明范围，并不限于上述技术特征的特定组合而成的技术方案，同时也应涵盖在不脱离上述发明构思的情况下，由上述技术特征或其等同特征进行任意组合而形成的其它技术方案。例如上述特征与本公开的实施例中公开的（但不限于）具有类似功能的技术特征进行互相替换而形成的技术方案。

Detailed Description

Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. While certain embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be embodied in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete. It should be understood that the drawings and embodiments of the present disclosure are for illustration purposes only and are not intended to limit the scope of the present disclosure.

It should be noted that, for convenience of description, only the portions related to the present invention are shown in the drawings. Embodiments of the present disclosure and features of embodiments may be combined with each other without conflict.

It should be noted that the terms "first," "second," and the like in this disclosure are merely used to distinguish between different devices, modules, or units and are not used to define an order or interdependence of functions performed by the devices, modules, or units.

It should be noted that references to "one", "a plurality" and "a plurality" in this disclosure are intended to be illustrative rather than limiting, and those of ordinary skill in the art will appreciate that "one or more" is intended to be understood as "one or more" unless the context clearly indicates otherwise.

The names of messages or information interacted between the various devices in the embodiments of the present disclosure are for illustrative purposes only and are not intended to limit the scope of such messages or information.

The present disclosure will be described in detail below with reference to the accompanying drawings in conjunction with embodiments.

Fig. 1 illustrates a flow 100 of some embodiments of an image cleaning method according to the present disclosure. The image cleaning method comprises the following steps:

And 101, controlling an acquisition vehicle to acquire data aiming at the speed limit plate so as to obtain a speed limit plate image sequence.

In some embodiments, the executing body (e.g., electronic device) of the above image cleaning method may control the collection vehicle to perform data collection for the speed limit plate, so as to obtain the speed limit plate image sequence. The limit board image data sequence may be an image sequence ordered according to the time sequence of the tray falling when the collecting vehicle collects the limit board. The landing can be to store the acquired speed limit sign image sequence to a magnetic disk. The time to drop may be the time the speed limit sign image is stored in the disk. The collection vehicle may be a vehicle that collects the speed limit sign. For example, the collection vehicle may be an unmanned vehicle.

In some optional implementations of some embodiments, the controlling the collection vehicle to collect data for the speed limit sign to obtain the speed limit sign image sequence may include the following steps:

and the first step is to control the acquisition vehicle to acquire data aiming at the speed limiting plate so as to obtain an acquisition image sequence. The image collection sequence may be an image sequence obtained by collecting the speed limit plate. The sequence of the image acquisition sequences can be a sequence obtained by sorting according to the time sequence of the tray.

And secondly, carrying out channel division on the acquired image sequences to obtain a plurality of channel image sequences. Wherein the channel images in the plurality of channel image sequences may be images including only one color channel. For example, the plurality of channel image sequences may include: a red channel image sequence, a green channel image sequence, and a blue channel image sequence.

Third, for each of the plurality of channel images in the sequence of channel images, performing the determining step of:

and a substep 1, performing smoothing processing on the channel image to obtain a smoothed image. The smoothing process may be an image in which high-frequency pixels in the channel image are removed. In practice, the execution subject may first convert the channel image into a frequency domain by fourier transform to obtain a frequency domain image. The frequency domain image may be an image obtained by fourier transforming a channel image located in a spatial domain to a frequency domain. The frequency domain can represent the change condition of the gray scale of the channel image. The spatial domain may be a two-dimensional planar coordinate system. Then, the frequency domain image is subjected to denoising processing by using a low-pass filter, and a smooth image sequence is obtained.

And 2, performing frequency domain conversion on the smooth image to obtain a spatial domain image. The spatial domain image may be an image in which pixels in the image are located in a two-dimensional plane coordinate system. In practice, the executing body may perform frequency domain conversion on the smoothed image by using inverse fourier transform, to obtain a spatial domain image.

And 3, determining the reflectivity of the spatial domain image. The reflectivity may be a ratio of the energy of the reflection of light by the object itself to the total energy of the light projected onto the object. As shown in fig. 2, the spatial domain image is an image composed of the product of incident light and reflectance. I (x, y) represents a spatial domain image. L (x, y) represents incident light. R (x, y) represents reflectance. The incident light may be external light when photographing the speed limit sign. In practice, the execution body may first perform normalization processing on the pixel values in the spatial domain image to obtain a normalized value. And then, carrying out Gaussian convolution filtering processing on the normalized value to obtain the reflectivity of the spatial domain image.

And fourthly, channel fusion is carried out on the obtained multiple reflectivity sequences, and a fused image sequence is obtained. The fused image sequence may be a speed-limiting plate image sequence obtained by fusing the corresponding reflectances of the three color channels.

And fifthly, carrying out equalization treatment on the fused image sequence to obtain a speed-limiting plate image sequence. In practice, the executing body may perform equalization processing on the fused image sequence by using a histogram equalization method, so as to obtain a speed-limiting plate image sequence.

The above technical solution and related content are taken as an invention point of the embodiments of the present disclosure, which solves the second "time-space domain only image filter or frequency domain image filter" technical problem mentioned in the background art, and performs filtering processing on collected image data in bad weather, so that edge information of an object in an image is unclear, and further, accuracy and safety of automatic driving vehicle recognition are low. Factors that lead to lower accuracy and safety of automatic driving vehicle identification tend to be as follows: and only the time-space domain image filter or the frequency domain image filter is used for filtering the acquired image data in severe weather, so that the edge information of an object in the image is unclear, and further the accuracy and safety of automatic driving vehicle identification are lower. If the above factors are solved, the effects of improving the definition of the edge information of the object in the image and reducing the overall brightness difference of the image can be achieved. To achieve this, the present disclosure first controls the above-described collection vehicle to perform data collection for the speed limit sign, resulting in a collection image sequence. Secondly, carrying out channel division on the acquired image sequences to obtain a plurality of channel image sequences. Here, the channel division can enhance the image to a fine granularity degree, and a clearer and noiseless image can be obtained. And then, carrying out smoothing processing on the channel image to obtain a smoothed image. Here, the smoothing process is advantageous in reducing noise in the image, and more highlighting important information in the image. And carrying out frequency domain conversion on the smooth image to obtain a spatial domain image. Here, the conversion from the frequency domain to the spatial domain is more advantageous for determining the reflectivity of the subsequent image. And determining the reflectivity of the spatial domain image. And then, carrying out channel fusion on the obtained multiple reflectivity sequences to obtain a fused image sequence. Here, determining the reflectivity may enhance defogging effects in bad weather, improving sharpness of the image and marginalizing information. And finally, carrying out equalization treatment on the fused image sequence to obtain the speed-limiting plate image sequence. Here, the equalization process may increase contrast and sharpness of the image, as well as reduce noise in the image. Therefore, the technical scheme combines the frequency domain and the spatial domain denoising method, so that the definition of images and the edge information of objects in an image sequence acquired in severe weather can be improved, and the accuracy and the safety of automatic driving vehicle identification are further improved.

Step 102, selecting a sample speed limit card image sequence from the speed limit card image sequence, and executing the following determining steps:

and 1021, sequentially inputting the sample speed limit plate image sequence into an initial target detection model to obtain a probability value sequence.

In some embodiments, the executing body may sequentially input the sample speed-limiting card image sequence to an initial target detection model to obtain a probability value sequence. The probability value in the probability value sequence represents the probability value of the speed limit plate in the sample speed limit plate image. The sample speed limit plate image sequence may be an image sequence for training the initial target detection model. The initial object detection model may include, but is not limited to: YOLO (You Only Look Once) model, SSD (Single Shot MultiBox Detector) model.

Step 1022, determining the target time point according to the probability value sequence.

In some embodiments, the execution body may determine the target time point according to the probability value sequence. The target time point may be a landing time point corresponding to a probability value located at an initial position, where the probability value is greater than or equal to a preset probability threshold in the probability value sequence. The initial position may be a first position.

As an example, the execution body may first compare each probability value in the probability value sequence with a preset probability threshold value to generate a comparison result, so as to obtain a comparison result sequence. And secondly, determining a comparison result which is larger than or equal to a preset probability threshold value in the comparison result sequence as a target result sequence. And finally, determining the acquisition time point corresponding to the probability value corresponding to the comparison result positioned at the initial position in the target result sequence as a target time point.

In some optional implementations of some embodiments, determining the target time point according to the probability value sequence may include the following steps:

the first step, probability values which are larger than or equal to the preset probability threshold value are screened out from the probability value sequence, and a target probability value sequence is obtained. Wherein the preset probability threshold may be 0.75.

And secondly, determining the sample speed limiting plate image sequence corresponding to the target probability value sequence as a target sample speed limiting plate image sequence.

And thirdly, determining an acquisition time point corresponding to the target sample speed limit plate image positioned at the initial position in the target sample speed limit plate image sequence as a target time point.

Step 1023, creating an initial variable window according to the target time point.

In some embodiments, the execution body may create an initial variable window according to the target time point. The initial variable window is used for determining a window of a sample speed limit plate image to be cleaned.

As an example, the execution subject may create an initial variable window for the sample speed limit board image corresponding to the target time point.

In some optional implementations of some embodiments, creating the initial variable window according to the target time point may include the steps of:

and determining a time point, of which the corresponding difference value is within a first preset time range, from the target time point as a starting time point. For example, the first preset time range may be 1 second to 3 seconds. The corresponding difference from the target time point may be 3 seconds.

And secondly, selecting a sample speed-limiting plate image corresponding to the same acquisition time point as the initial time point from the sample speed-limiting plate image sequence as an initial sample speed-limiting plate image. The collection time point can be a time point when the collection vehicle collects the speed limit plate and falls into the memory.

And thirdly, determining a time point, of which the corresponding difference value between the target time point and the target time point is within a second preset time range, as a termination time point. For example, the second preset time range may be 1 second to 2 seconds. The corresponding difference from the above-mentioned point in time may be 1 second.

And step four, selecting a sample speed limit plate image corresponding to the same acquisition time point as the termination time point from the sample speed limit plate image sequence as a termination sample speed limit plate image.

And fifthly, determining a sample speed limit plate image which is positioned between the initial sample speed limit plate image and the final sample speed limit plate image in the sample speed limit plate image sequence as a sample speed limit plate image sequence included in the initial variable window.

Step 1024, generating at least one image sequence according to the initial variable window.

In some embodiments, the execution body may generate at least one image sequence according to the initial variable window. Wherein the at least one image sequence may be a sample speed limit card image sequence included in the at least one initial variable window.

As an example, the execution subject may create an initial variable window for a probability value greater than or equal to a preset probability threshold value from probability values corresponding to the sample speed limit card image sequence, to obtain at least one image sequence.

In some optional implementations of some embodiments, generating at least one image sequence according to the initial variable window may include:

first, based on a first sequence of probability values, the following sequence determination steps are performed:

and 1, in response to determining that the initial variable window does not comprise a sample speed limit plate image sequence corresponding to the acquisition time point which is the same as the sequence termination time point, expanding the initial variable window to obtain an expanded variable window and an expanded window. The extended window may be a window corresponding to a portion of the extended variable window larger than the initial variable window. The extended variable window may be a window obtained by extending a time point corresponding to a termination position of the initial variable window. The first probability value sequence is the probability value sequence from the target time point to a sequence termination time point, and the sequence termination time point may be a time point corresponding to a probability value located at a termination position of the probability value sequence.

And 2, in response to determining that no probability value which is larger than or equal to the preset probability threshold exists in the probability value sequence corresponding to the sample speed limit plate image sequence included in the expansion window, adding the sample speed limit plate image sequence included in the expansion variable window to at least one preset image sequence. The at least one preset image sequence may be a preset image sequence including at least one image sequence.

And 3, in response to determining that a second probability value larger than or equal to the preset probability threshold exists in the second probability value sequence, screening the second probability value larger than or equal to the preset probability value from the second probability value sequence, and obtaining a target screening probability value sequence. The second probability value sequence is a probability value sequence located between the probability value sequence and a probability value corresponding to the sequence termination time point from the extended probability value. The expansion probability value is a probability value corresponding to a sample speed limit plate image sequence positioned at a termination position in the obtained expansion variable window.

And step 4, aiming at the sample speed limiting plate image corresponding to the probability value positioned at the initial position in the target screening probability value sequence, creating a sample variable window. The sample variable window can be used for determining a window of a sample speed limit plate image to be cleaned.

And a substep 5, determining the at least one preset image sequence as the at least one image sequence in response to determining that the sample variable window comprises a sample speed limit plate image corresponding to the same acquisition time point as the sequence termination time point.

And a second step of determining a probability value sequence located between the window termination time point and the sequence termination time point as a first probability value sequence in response to determining that the sample variable window does not include a sample speed limit plate image corresponding to the same acquisition time point as the sequence termination time point, and performing the sequence determining step again. The window ending time point is a time point corresponding to the sample speed limit plate image at the ending position of the sample variable window.

In some optional implementations of some embodiments, the expanding the initial variable window to obtain an expanded variable window and an expanded window includes:

first, based on the initial variable window, the following expansion steps are performed:

and 1, determining probability values corresponding to the sample speed limit plate images between the target time point and the ending time point as a window probability value sequence. The window probability value sequence may be a probability value corresponding to a sample speed limit plate image sequence included between a target time point and a termination time point.

And 2, determining at least one window probability value which is larger than or equal to the preset probability threshold value as a target window probability value sequence in response to determining that the window probability value which is larger than or equal to the preset probability threshold value exists in the window probability value sequence. The target window probability value sequence may be a probability value sequence that is located in the window probability value sequence and is equal to or greater than the preset probability threshold.

And 3, determining a target window probability value at the ending position of the target window probability value sequence as a target ending window probability value.

And step 4, determining the acquisition time point of the target termination window probability value as an adjustment time point.

And step 5, expanding the initial variable window according to the adjustment time point to obtain an expanded variable window and the expanded window. The extended variable window may be a window extended backward by a preset time at a time point corresponding to a termination position of the initial variable window. For example, the preset time may be 1 second.

As an example, the execution body expands the initial variable window for 1 second from the adjustment time point to the start time point, and then expands the initial variable window for 1 second.

And a substep 6, wherein the expanding step is ended in response to determining that no probability value larger than or equal to the preset probability threshold exists in the probability value sequence corresponding to the sample speed limit plate image sequence included in the expanding window.

Optionally, the above execution body may further execute the following steps:

and in response to determining that a probability value greater than or equal to the preset probability threshold exists in a probability value sequence corresponding to a sample speed-limiting card image sequence included in the expansion window, determining a collection time point of the sample speed-limiting card image positioned at a termination position in the initial variable window as a target time point, determining the expansion variable window as the initial variable window, and executing the expansion step again.

Optionally, after the expanding the initial variable window in response to determining that the initial variable window does not include the sample speed limit plate image sequence corresponding to the same acquisition time point as the sequence termination time point, the method further includes:

and in response to determining that the initial variable window comprises a sample speed limit plate image sequence corresponding to the same acquisition time point as the sequence termination time point, adding the sample speed limit plate image sequence comprising the initial variable window to the at least one preset image sequence.

Step 1025, cleaning at least one image sequence to obtain a processed speed limit plate image sequence.

In some embodiments, the executing body may perform a cleaning process on the at least one image sequence to obtain a processed speed-limiting card image sequence. The processed speed limit plate image sequence can be an image sequence with a probability threshold smaller than a preset probability threshold and including a speed limit plate. The above-described cleaning process may include, but is not limited to: and (3) performing cleaning treatment and manual detection cleaning treatment on at least one image sequence by using a statistical method.

Step 1026, determining a ratio of the number of probability values in the probability value sequence that is greater than or equal to the preset probability threshold to the number of probability values in the probability value sequence.

In some embodiments, the execution body may determine a ratio of a number of probability values in the probability value sequence that is greater than or equal to a preset probability threshold to a number of probability values in the probability value sequence. The preset probability threshold may represent that training of the initial target detection model is completed. For example, the preset probability threshold may be 0.75.

In step 1027, in response to determining that the ratio is greater than or equal to the preset threshold, the initial variable window is determined to be the target variable window.

In some embodiments, the executing entity may determine the initial variable window as the target variable window in response to determining that the ratio is greater than or equal to a preset threshold. Wherein, the target variable window may be a window corresponding to the initial target detection model. The target variable window may be a window in which the length of the variable window decreases as the accuracy of the initial target model increases.

And step 103, in response to determining that the ratio is smaller than the preset threshold, adjusting the initial variable window and the initial target detection model, taking the processed speed-limiting plate image sequence as a sample speed-limiting plate image sequence, respectively determining the adjusted initial variable window and the adjusted initial target detection model as an initial variable window and an initial target detection model, and executing the determining step again.

In some embodiments, the executing body may adjust the initial variable window and the initial target detection model in response to determining that the ratio is smaller than the preset threshold, determine the adjusted initial variable window and the adjusted initial target detection model as the initial variable window and the initial target detection model, respectively, using the processed speed limit image sequence as the sample speed limit image sequence, and execute the determining step again.

The above embodiments of the present disclosure have the following advantages: the image cleaning method of some embodiments of the present disclosure can reduce the omission ratio of image cleaning and improve the efficiency of image cleaning by using a variable window with adjustable size and a target detection model. Specifically, the reason for the reduced safety of the related autopilot is that: the efficiency of manual cleaning is lower, and cleaning cycle and cleaning cost are higher, and the accuracy of utilizing the target detection model trained in advance to detect speed limit plate is lower, and the omission ratio of the image to some marginal scenes is higher, leads to the image cleaning efficiency lower. Based on this, the image cleaning method of some embodiments of the present disclosure may first control the collection vehicle to perform data collection for the speed limit sign, resulting in a speed limit sign image sequence. Here, the speed limit tile image sequence facilitates reducing miss rate in combination with the initial variable window and subsequent adjustment of the variable window and the target detection model. Then, selecting a sample speed limit plate image sequence from the speed limit plate image sequences, and executing the following determining steps: and sequentially inputting the sample speed-limiting plate image sequence to an initial target detection model to obtain a probability value sequence, wherein the probability value in the probability value sequence represents the probability value of the speed-limiting plate in the sample speed-limiting plate image. Here, the resulting sequence of probability values facilitates a subsequent determination of the accuracy of the initial target detection model. And determining a target time point according to the probability value sequence. Here, obtaining the target time point facilitates creating an initial variable window, and reducing the omission ratio of the sample speed limit plate image sequence cleaning, and obtaining the edge image data of the speed limit plate facilitates improving the model detection rate. And creating an initial variable window according to the target time point, wherein the initial variable window is used for determining a window of the sample speed limit plate image to be cleaned. Here, creating an initial variable window can acquire an edge image in the sample speed-limiting card image sequence, and improve the omission ratio of the sample speed-limiting card image sequence cleaning. At least one image sequence is generated based on the initial variable window. Here, the acquisition of at least one image sequence facilitates the subsequent cleaning process, resulting in an edge image. And then, cleaning the at least one image sequence to obtain a processed speed-limiting plate image sequence. Here, the cleaning of the at least one image sequence can improve the detection accuracy of the target detection model, and reduce the cleaning period and the cleaning cost of the sample speed-limiting plate image sequence. Determining the ratio of the number of probability values in the probability value sequence which are larger than or equal to a preset probability threshold value to the number of probability values in the probability value sequence; and in response to determining that the ratio is greater than or equal to a preset threshold, determining the initial variable window as a target variable window. And finally, in response to determining that the ratio is smaller than the preset threshold, adjusting the initial variable window and the initial target detection model, taking the processed speed-limiting plate image sequence as a sample speed-limiting plate image sequence, respectively determining the adjusted initial variable window and the adjusted initial target detection model as an initial variable window and an initial target detection model, and executing the determining step again. Here, the omission ratio of the sample speed limit plate image sequence cleaning can be reduced based on the target detection model and the variable window, and the initial variable window and the initial target detection model can be adjusted to improve the efficiency of model and image cleaning and reduce the cleaning period and the cleaning cost. Therefore, the image cleaning method can reduce the omission ratio of image cleaning and improve the efficiency of image cleaning and the detection efficiency of the target detection model by using the variable window with adjustable size and the target detection model.

With further reference to fig. 3, as an implementation of the method shown in the above figures, the present disclosure provides embodiments of an image cleaning apparatus, which correspond to those method embodiments shown in fig. 1, which may find particular application in a variety of electronic devices.

As shown in fig. 3, an image cleaning apparatus 300 includes: an acquisition unit 301, an execution unit 302 and an adjustment unit 303. Wherein the acquisition unit 301 is configured to: and controlling the acquisition vehicle to acquire data aiming at the speed limit plate so as to obtain a speed limit plate image sequence. The execution unit 302 is configured to: selecting a sample speed limit plate image sequence from the speed limit plate image sequences, and executing the following determining steps: sequentially inputting the sample speed-limiting plate image sequence into an initial target detection model to obtain a probability value sequence, wherein the probability value in the probability value sequence represents the probability value of the speed-limiting plate in the sample speed-limiting plate image; determining a target time point according to the probability value sequence; creating an initial variable window according to the target time point, wherein the initial variable window is used for determining a window of a sample speed limit plate image to be cleaned; generating at least one image sequence according to the initial variable window; cleaning the at least one image sequence to obtain a processed speed limit plate image sequence; determining the ratio of the number of probability values in the probability value sequence which are larger than or equal to a preset probability threshold value to the number of probability values in the probability value sequence; and in response to determining that the ratio is greater than or equal to a preset threshold, determining the initial variable window as a target variable window. The adjustment unit 303 is configured to: and in response to determining that the ratio is smaller than the preset threshold, adjusting the initial variable window and the initial target detection model, taking the processed speed-limiting plate image sequence as a sample speed-limiting plate image sequence, respectively determining the adjusted initial variable window and the adjusted initial target detection model as the initial variable window and the initial target detection model, and executing the determining step again.

It will be appreciated that the elements described in the image cleaning apparatus 300 correspond to the various steps in the method described with reference to fig. 1. Thus, the operations, features and advantages described above with respect to the method are equally applicable to the image cleaning apparatus 300 and the units contained therein, and are not described herein.

Referring now to fig. 4, a schematic diagram of an electronic device (e.g., electronic device) 400 suitable for use in implementing some embodiments of the present disclosure is shown. The electronic device shown in fig. 4 is merely an example and should not impose any limitations on the functionality and scope of use of embodiments of the present disclosure.

As shown in fig. 4, the electronic device 400 may include a processing means (e.g., a central processing unit, a graphics processor, etc.) 401, which may perform various suitable actions and processes according to a program stored in a Read Only Memory (ROM) 402 or a program loaded from a storage means 408 into a Random Access Memory (RAM) 403. In the RAM403, various programs and data necessary for the operation of the electronic device 400 are also stored. The processing device 401, the ROM 402, and the RAM403 are connected to each other by a bus 404. An input/output (I/O) interface 305 is also connected to bus 404.

In general, the following devices may be connected to the I/O interface 405: input devices 406 including, for example, a touch screen, touchpad, keyboard, mouse, camera, microphone, accelerometer, gyroscope, etc.; an output device 407 including, for example, a Liquid Crystal Display (LCD), a speaker, a vibrator, and the like; storage 408 including, for example, magnetic tape, hard disk, etc.; and a communication device 409. The communication means 409 may allow the electronic device 400 to communicate with other devices wirelessly or by wire to exchange data. While fig. 4 shows an electronic device 400 having various means, it is to be understood that not all of the illustrated means are required to be implemented or provided. More or fewer devices may be implemented or provided instead. Each block shown in fig. 4 may represent one device or a plurality of devices as needed.

In particular, according to some embodiments of the present disclosure, the processes described above with reference to flowcharts may be implemented as computer software programs. For example, some embodiments of the present disclosure include a computer program product comprising a computer program embodied on a computer readable medium, the computer program comprising program code for performing the method shown in the flow chart. In such embodiments, the computer program may be downloaded and installed from a network via communications device 409, or from storage 408, or from ROM 402. The above-described functions defined in the methods of some embodiments of the present disclosure are performed when the computer program is executed by the processing device 401.

It should be noted that, in some embodiments of the present disclosure, the computer readable medium may be a computer readable signal medium or a computer readable storage medium, or any combination of the two. The computer readable storage medium can be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or a combination of any of the foregoing. More specific examples of the computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In some embodiments of the present disclosure, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. In some embodiments of the present disclosure, however, the computer-readable signal medium may comprise a data signal propagated in baseband or as part of a carrier wave, with the computer-readable program code embodied therein. Such a propagated data signal may take any of a variety of forms, including, but not limited to, electro-magnetic, optical, or any suitable combination of the foregoing. A computer readable signal medium may also be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to: electrical wires, fiber optic cables, RF (radio frequency), and the like, or any suitable combination of the foregoing.

In some implementations, the clients, servers may communicate using any currently known or future developed network protocol, such as HTTP (Hyper Text Transfer Protocol ), and may be interconnected with any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network ("LAN"), a wide area network ("WAN"), the internet (e.g., the internet), and peer-to-peer networks (e.g., ad hoc peer-to-peer networks), as well as any currently known or future developed networks.

The computer readable medium may be contained in the electronic device; or may exist alone without being incorporated into the electronic device. The computer readable medium carries one or more programs which, when executed by the electronic device, cause the electronic device to: controlling an acquisition vehicle to acquire data aiming at the speed limit plate to obtain a speed limit plate image sequence; selecting a sample speed limit plate image sequence from the speed limit plate image sequences, and executing the following determining steps: sequentially inputting the sample speed-limiting plate image sequence into an initial target detection model to obtain a probability value sequence, wherein the probability value in the probability value sequence represents the probability value of the speed-limiting plate in the sample speed-limiting plate image; determining a target time point according to the probability value sequence; creating an initial variable window according to the target time point, wherein the initial variable window is used for determining a window of a sample speed limit plate image to be cleaned; generating at least one image sequence according to the initial variable window; cleaning the at least one image sequence to obtain a processed speed limit plate image sequence; determining the ratio of the number of probability values in the probability value sequence which are larger than or equal to a preset probability threshold value to the number of probability values in the probability value sequence; in response to determining that the ratio is greater than or equal to a preset threshold, determining the initial variable window as a target variable window; and in response to determining that the ratio is smaller than the preset threshold, adjusting the initial variable window and the initial target detection model, taking the processed speed-limiting plate image sequence as a sample speed-limiting plate image sequence, respectively determining the adjusted initial variable window and the adjusted initial target detection model as the initial variable window and the initial target detection model, and executing the determining step again.

Computer program code for carrying out operations for some embodiments of the present disclosure may be written in one or more programming languages, including an object oriented programming language such as Java, smalltalk, C ++ and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The program code may execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any kind of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or may be connected to an external computer (for example, through the Internet using an Internet service provider).

The flowcharts and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.

The units described in some embodiments of the present disclosure may be implemented by means of software, or may be implemented by means of hardware. The described units may also be provided in a processor, for example, described as: a processor includes an acquisition unit, an execution unit, and an adjustment unit. The names of these units do not in any way limit the units themselves, for example, the acquisition unit may also be described as "a unit that controls the acquisition vehicle to perform data acquisition for the speed limit sign, resulting in a sequence of speed limit sign images".

The functions described above herein may be performed, at least in part, by one or more hardware logic components. For example, without limitation, exemplary types of hardware logic components that may be used include: a Field Programmable Gate Array (FPGA), an Application Specific Integrated Circuit (ASIC), an Application Specific Standard Product (ASSP), a system on a chip (SOC), a Complex Programmable Logic Device (CPLD), and the like.

The foregoing description is only of the preferred embodiments of the present disclosure and description of the principles of the technology being employed. It will be appreciated by those skilled in the art that the scope of the invention in the embodiments of the present disclosure is not limited to the specific combination of the above technical features, but encompasses other technical features formed by any combination of the above technical features or their equivalents without departing from the spirit of the invention. Such as the above-described features, are mutually substituted with (but not limited to) the features having similar functions disclosed in the embodiments of the present disclosure.