恶劣天气下的车牌信息识别

License Plate Information Recognition Algorithm under Severe Weather

车牌信息识别包括图像预处理、车牌定位、字符分割、字符识别等步骤。在有雾霾、扬沙的恶劣天气状况下,由于户外能见度较差,造成采集的车牌图像模糊、颜色保真度差,降低了车牌检测和识别的准确性和可靠性。针对这一问题,本文利用图像饱和度估计介质传输的方法对图像去雾,同时使用白平衡的方法来进一步消除黄色或细粉尘对图像色彩造成的偏差。这种方法有效还原了车辆图像原本的信息特征,并且在去雾效率和计算复杂度两方面都表现出了优越的性能,为后续车牌信息的识别奠定了基础。

License plate information recognition includes image preprocessing,license plate location,character segmentation,character recognition and other steps. Under severe weather conditions with haze and blowing sand,poor outdoor visibility results in blurred license plate images and poor color fidelity,which reduces the accuracy and reliability of license plate detection and recognition. In response to this problem,this paper uses the method of image saturation to estimate the medium transmission to defog the image,and at the same time uses the white balance method to further eliminate the deviation of the image color caused by yellow or fine dust. This method effectively restores the original information characteristics of the vehicle image,and shows superior performance in both dehazing efficiency and computational complexity,laying a foundation for subsequent license plate information recognition.