A New Vehicle Distance Measure Method Based on the License Plate with Image Edge Detection and Color Recognition in the Complex Environment

License plate detection and vehicle distance measurement are important in applying modern computer image processing in the transportation field. This paper mainly processes Chinese license plates in natural scenes by traditional image processing algorithms and applies them in real scenes. We measured the distance between motor vehicles with the constant size of the license plate by derived the distance between motor vehicles by combining the distance ratio in the real world and photograph. More specifically, we used edge detection to remove some smooth background, including blue lumps that will first influence the detection of license plate. Based on the result gotten in edge detection, we used color detection to eliminate other useless edges that are not blue. To restrain the influence of the blue area in the margins and the dim brightness in the whole image, we constructed a blue color distance matrix to find the bluest color closed to the center and used this color as a new standard to construct another new blue color distance. Setting a thresh, we binarize this image to find blue areas and use the minimum bounding box to circle each connection domain. By find the ratio and the distance of each minimum bounding box to the center of the image, we give a grade to each bounding box and find the target connection domain. Then, the least square method error correction was employed to detect the license plate and measure the distance between vehicles. At last, we sampled results randomly to correct errors and get results with 0.0428 mean square error, which is much less than 0.4422, which is also the mean square error before the least square method correction. These results reflected the algorithm also gets acceptable results compared to other methods.