Research On Vehicle Dimensions Measuring System Under The Machine Vision

Since the first car was born in1886, with the progress of mechanical, electronicand age, the number of cars has increased rapidly. This brings up a series of seriousproblems for road traffic management, leading to serious social problems. Moderntraffic management is more and more troubled by frequent traffic accidents, severeroad damage, environmental pollution, traffic congestion and other problems. Ourcountry began to enforce the standard law of "road vehicle dimensions, axle loadand quality limits"(GB1586-2004) since October1,2004. The Purpose is to increasethe legal supervision of the vehicle modification. So the research of Maintenancedetection technology has a very important practical significance and theoreticalresearch value.In order to satisfy the demand of modern intelligent traffic management, in viewof the freight car transfinite problem in intelligent traffic, this paper uses the machinevision technology to study the outline of a suitable for operating freightmeasuring characteristics and size parameters extraction and non-contactmeasurement method, bases on machine vision technology which builds the generalprocess of freight car size parameter feature extraction system, has reached thevehicle types of effective recognition and the purpose of the digital measurement ofsize parameters of the vehicle.This paper firstly uses the principle of background subtraction automotiveimages with complex backgrounds to achieve positioning; Secondly, in order toovercome the weaknesses undetected problems which exist in the traditional Cannyedge image edge detection algorithms, by the combination of multiple OTSUthreshold method and image block theory,to achieve the effective detection andextraction of auto edge profile. After that, using the principle of binocular visionresearch a horizon calibration method based on the coordinate transformation, in order to realize the vehicle size parameter calibration, and verify the feasibility of thecalibration method.Finally, the paper studies the system PC management platformdesign methodology and other related issues, then designs and develops a kind of PCmanagement of a friendly human-machine interface.The research contents and methods of this article is a kind of Supplementaryway of technical means in the intelligent traffic management field, is a useful attemptof using technological means to achieve intelligent traffic management.