Design Of Vehicle Obstacle Recognition And Distance Measurement System Based On Machine Vision

A front car obstacle recognition and distance detection system based on machine vision is designed in this paper.The system is a kind of vehicle active safety protection system,it is an important part of the intelligent network environment perception system in automobile.At the same time,obstacle recognition and ranging is also one of the most fundamental and core technologies of intelligent driving.Mastering and perfecting this technology can not only push forward the development of intelligent network vehicles,but also can reduce the incidence of rear end collision accidents and reduce casualties and the losses of property,which is of great practical significance.This system is built on base of DM6437 as hardware platform,CCS as programming environment,collecting road information by camera,processing and analyzing the input images,so as to recognize obstacles in front of vehicles.Some main methods of image processing is used to deal with the image which the camera put in through grayscale,filtering,ROI selection,binarization edge detection and image processing operations into the image information we need.We frame to select the target and display it on the display.The distance measurement should be carried out after the obstacle is detected.We establishes a safe distance model first according to the definition of the safe distance.Then an obstacle location model is established through the analysis of the different coordinate system.After calculating the relevant parameters according to the specific numerical camera installed in the car,convert the value through the two coordinates and calculate the distance,then,display the value of the distance.A real vehicle road experiment is carried out in order to check the practicability and the actual working situation of the system.We use a laser ranging system to compare the ranging results with this system,then,test and analyze the data.The experiment proved the practicability of the algorithm which is used in the system,and through the analysis of the experimental results,a perfect scheme for the system was put forward.