A video-based vehicle detection method

Abstract

The present invention provides a vehicle detection method based on video, which has two detection modes: daytime and nighttime detection modes: during the daytime, motion detection combined with license plate detection is used, and at night, vehicle light detection combined with license plate detection is used, that is, the license plate detection result is used as Whether it is the primary judgment criterion for a car, if no license plate is detected, it will be supplemented with the vehicle detected by motion during the day, and the detected car lights at night. The license plate detection is carried out within the moving target, which reduces the scope of test. The invention provides a video-based vehicle detection method with high detection accuracy, strong adaptability and hardware resource saving.

Description

technical field [0001] The invention relates to the field of intelligent transportation, in particular to a video-based vehicle detection method. Background technique [0002] The rapid development of my country's urban road traffic has brought great convenience to people's daily life. At the same time, the accompanying negative effects have become increasingly prominent, such as traffic congestion and frequent traffic accidents. To solve such problems, intelligent transportation systems have emerged. The video-based vehicle detection system is an important part of intelligent transportation and a prerequisite for subsequent vehicle behavior analysis and judgment, and its importance is self-evident. The problems of high actual vehicle density, complex lighting conditions, large amount of image data, and real-time performance put forward higher requirements for the design of detection methods. High precision, real-time, high efficiency, and stability have become the basic ch...

AI Technical Summary

Problems solved by technology

[0003] At present, there are many methods for video-based vehicle detection, but these methods have great limitations, such as: 1. The illumination of the actual scene is complex and changeable, resulting in a decrease in detection accuracy; 2. The phenomenon of adhesion is serious, resulting in missed detection and false detection 3. The image resolution is getting higher and higher, and the resulting data volume has increased significantly, resulting in a single frame image processing time is too long, CPU usage and memory usage are high, so it cannot meet the actual needs

Images