自然环境下，汽车图像背景复杂、光照不均匀，如何在自然背景中准确地确定牌照区域是整个识别过程的关键。首先对采集到的视频图像进行大范围相关搜索，找到符合汽车牌照特征的若干区域作为候选区，然后对这些侯选区域做进一步分析、评判，最后选定一个最佳的区域作为牌照区域，并将其从图象中分割出来。
In the natural environment, the background of vehicle image is complex and the illumination is uneven. How to accurately determine the license plate area in the natural background is the key of the whole recognition process. Firstly, the video image is searched in a large range to find some regions which meet the characteristics of vehicle license plate as candidate regions. Then, the candidate regions are further analyzed and evaluated. Finally, the best region is selected as license plate region, and it is segmented from the image.
牌照字符分割
License plate character segmentation
完成牌照区域的定位后，再将牌照区域分割成单个字符，然后进行识别。字符分割一般采用垂直投影法。由于字符在垂直方向上的投影必然在字符间或字符内的间隙处取得局部最小值的附近，并且这个位置应满足牌照的字符书写格式、字符、尺寸限制和一些其他条件。利用垂直投影法对复杂环境下的汽车图像中的字符分割有较好的效果。
After the location of license plate area is completed, the license plate area is divided into single characters, and then recognition is carried out. Character segmentation generally uses vertical projection method. Because the projection of characters in the vertical direction must obtain the local minimum value near the gap between characters or within characters, and this position should meet the character writing format, character, size limit and some other conditions of license plate. The method of vertical projection has a good effect on character segmentation of automobile image in complex environment.