A license plate location method

Abstract

The invention discloses a license plate positioning method, and belongs to the technical field of image processing. The license plate positioning method comprises the steps of preprocessing the images of a license plate to be positioned to obtain a plurality of candidate rectangular coordinates, initializing a sliding window H on the basis of the current candidate rectangular coordinates, initializing the color of the frame of H so that the color of the frame of H can be the color distributed in most of the regions, and enabling H to continuously shrink according to the energy minimization principle and carrying out angle variation on the images in rectangular regions till the rectangle with the minimum energy is obtained. According to the method, because the obtained boundary region of the license plate is very close to the first character and the last character, processing complexity of segmentation of the characters of the license plate is effectively reduced, and the efficiency and the robustness of license plate recognition are effectively improved. Meanwhile, some special license plates such as double-row license plates and the license plates of armed police are better positioned and corrected.

Description

technical field [0001] The invention belongs to the technical field of image processing, in particular to the positioning of license plates in images. Background technique [0002] In recent years, with the continuous growth of the domestic economy, the growth rate of motor vehicles in the country has been maintained at more than 15%, which has increased the difficulty of vehicle management in various industries related to vehicles, making it necessary to increase manpower input and Handle issues such as efficiency. At the same time, the application of computer vision technology in the field of image processing has gradually transformed vehicle management from a manual method to an intelligent and fully automatic method, which can effectively solve the current problems. License plate recognition technology (character recognition of license plate images) has great application value in the field of public safety, traffic management, gas stations and parking lots. License pla...

AI Technical Summary

Problems solved by technology

[0004] (1) Based on the edge detection method, the edge detection algorithm is the most basic and effective among these three types of algorithms. For simple scenes, such as toll stations, etc., a higher detection rate can be achieved, but for areas with complex scenes , such as streets, etc., the false detection rate will be relatively high, making it impossible to accurately locate the position of the license plate, and because the complexity of the scene is relatively high, there are many edges of color jumps in the scene, which makes a lot of work to be done in the process of license plate positioning Useless work reduces the utilization of the CPU, which will inevitably lead to an unsatisfactory real-time experience

[0005] (2) The color-based license plate positioning method is essentially a kind of edge-based algorithm. It uses the edge of the color image and the internal color information to locate it. Therefore, the resolution of the picture is relatively high. If the picture If it is blurry, or in night, rainy days, etc., the success rate of license plate location will be greatly reduced, and it does not have good robustness

[0006] (3) The positioning method based on machine learning, first of all, there must be enough training samples, and find good features and training methods. Many people use features to locate license plates based on haar transformation and adaboost, which can achieve relatively high The detection rate is high, but the detection range is large and the false detection rate is relatively high