Characteristic area positioning method, car body color depth and color recognition method

Abstract

The invention discloses a method for positioning a characteristic region representing a dark and light color and a color of a vehicle body, which comprises the steps of: determining a vehicle region according to a reference region of a vehicle in an acquired vehicle image; constructing an energy distribution function according to texture characteristics and structural characteristics of the vehicle region; and positioning the characteristic region used for representing the dark and light color and the color of the vehicle body by taking a point with maximum energy in the energy distribution function as a central point. The invention also provides a method for identifying the dark and light color of the vehicle body and a method for identifying the color of the vehicle body. The application of the method can reduce computation amount, quicken computation speed, and improve identification accuracy rate.

Description

technical field [0001] The present invention relates to the technical field of image processing, in particular, to a method for locating vehicle body shades and characteristic regions of colors, a method for identifying vehicle body shades, and a method for identifying vehicle body colors. Background technique [0002] In places such as security checkpoints and expressway toll stations, the intelligent traffic monitoring system can be used to monitor the vehicles on the road. Accurate vehicle-related information (such as license plate number, model characteristics, vehicle brand, or body color and color, etc.) is obtained from vehicle images. Among them, license plate recognition technology has developed relatively maturely through research and development in recent years. For example, the application number is 200410054049 The Chinese patent application published in "Automobile License Plate Recognition Device with Lane Panoramic Monitoring Function and Recognition Method" ...

AI Technical Summary

Problems solved by technology

[0005] However, the above-mentioned car body color recognition method uses an algorithm based on dynamic images. Compared with algorithms based on static images, dynamic algorithms require a large amount of calculation, and the accuracy of the obtained recognition results is relatively poor.

In addition, the above-mentioned vehicle body color recognition method is to perform color recognition on all pixels in the entire vehicle video image after the background is removed, which not only requires a large amount of calculation, but also introduces areas in the vehicle image that cannot be used to represent the color of the vehicle body, such as windows, etc. resulting in poor reliability and low recognition accuracy

In order to improve the reliability, it is necessary to additionally use the D-S evidence theory method based on artificial statistical experience to eliminate the interference of the color of the wheels, lights, windows and other positions on the body color recognition, which is easy to affect the versatility of the recognition system and increase the complexity of the system

Images