Vehicle detecting and early-warning method based on machine vision and cascade classifier

Abstract

The invention provides a vehicle detecting and early-warning method based on machine vision and a cascade classifier, which comprises the steps of carrying out cascade classifier training on positive and negative samples by using edited automobile pictures of known automobile models on the market; building a mathematical model of the recognition frame width and the real distance through a lot of measured data; acquiring images in real time; carrying out image preprocessing, and framing a region-of-interest; carrying out target recognition on the region-of-interest; performing data analysis on a recognition result so as to screen and track a target; carrying out recognition frame moving average filtering processing on the tracked target; and calculating the real distance of the recognition target, calculating the moving speed of a vehicle through the frame difference time and the real distance difference so as to acquire the time when collision possibly occurs and display early-warning information. The method provided by the invention effectively improves the recognition efficiency for a moving target in the real environment and enhances the capacity for detecting and tracking the moving target in a complex environment through the processes of cascade classifier training performed by using the positive and negative samples, real-time image acquisition, image preprocessing, target recognition, target tracking and screening, target early-warning and the like.

Description

technical field [0001] The invention relates to the technical field of intelligent measurement and calculation of vehicle distances, in particular to a vehicle detection and early warning method based on machine vision and cascade classifiers. Background technique [0002] At present, during the driving process, due to driving fatigue and other reasons, drivers often need an effective tool for collision warning during fast driving. To perform collision warning, the problem of real-time distance measurement must first be solved. Currently, the commonly used distance measurement method There are four types: ultrasonic ranging, millimeter-wave radar ranging, laser ranging, and camera system ranging. [0003] Ultrasonic ranging refers to the method of calculating the distance of the target by using the time difference between transmitting and receiving ultrasonic waves. The ultrasonic rangefinder has the advantages of simple principle, convenient manufacture and relatively low ...

AI Technical Summary

Problems solved by technology

[0007] The above methods have disadvantages such as poor accuracy, high cost, single function, need to install expensive hardware equipment, slow imaging speed, and are greatly affected by the environment. They are not suitable for long-distance real-time speed measurement and real-time collision warning. Therefore, At present, drivers are in urgent need of an easy-to-use and effective tool for collision warning during fast driving.

Images