Study On Testing System Of The Position Deviation Of Vehicle-body And Axle Based On Stereo Vision

With rapid increase of the preserving amount of cars and the rapid development of the vehicle detection, the on-line automobile detection is towards automatization and intellectualization. Although the majority of the project has achieved the automation, the inferior projects are still in the stage of relying on manual testing, which has seriously affected the speed of vehicles detection, hinder the development of intellectualization and automation. As a result, it is imperative to use advanced computer technology and automatic test equipment to replace the traditional detection methods .On the other hand, from the national standards, it can also be seen that these inferior projects have been paid increasing attention to, which require more and more high-precision measurements. According to multiple performance requirement and detecting methods for commercial vehicle (National Standards) GB18565-2001, the equipment requirements for vehicle goes: vehicle equipment should be complete, intact, effective and fastening parts should be well-connected. The body should be modified, height difference must be no larger than 40 mm of the symmetrical body parts of the outer edge (from 1.5 m high with measurement); wheelbase defect between the right and the left should be less than 1.5/1000 of the wheelbase. According to " Specifications for multiple-function detecting test station of vehicle transport "(GB/T 17993-2005), which is issued by General Administration of Quality Supervision,Inspection and Quarantine of the People's Republic of China (AQSIQ) and the Standardization Administration of China on July 21, 2005, and is implemented on December 1, 2005, the rules for the high difference between symmetry around the outer edge of the body and the body as well as the wheelbase are as follows: measuring range: 0 ~ 5000mm (height difference), 0 ~ 20000mm (wheelbase defect); discrimination 1mm; error allowed±0.15mm. For Such testing precision, the traditional method of testing is no longer able to meet the requirements.Based on background mentioned above, this subject explores automation detection system in order to study the deviation about vehicle body and axle to replace the traditional manual testing method, achieving to measure the height difference of symmetry points and angle of shaft declination fast and accurately. Considering the detection technology characteristics of the height difference of symmetrical points and the angle of shaft declination, this paper designs and develops the detection system based on the stereovision. By using the binocular stereo vision, the automation detection about the height difference of symmetrical points and the angle of shaft declination is realized. Based on the theory of stereo vision, starting with the extraction and matching of image feature, camera calibration and other key technologies, this paper shows a algorithm which is suitable for the automation detection system mentioned in this paper. Through the experimental research, we have achieved a number of innovative research results. The main contents of the study and the results are as follows:1.Study the background and meanings of the paper. The paper summarized the actuality of testing techniques of the Position Deviation of Vehicle-body and Axle internal and overseas.2.Research and improve visual image feature extraction and matching algorithms. The key technologies of three-dimensional vision are discussed from the edge detection, and the line feature extraction, matching three-dimensional detail. In this paper, the traditional algorithm is analyzed and the improved algorithm is proposed to achieve the extraction of edge, corner and the straight-line accurately. The main assignment of this paper includes that having studied the matching technology in the stereo vision and proposed the stereo matching method based on parallax positioning bound. This paper predigests the matching matrix by using the polar of geometric constraints, the combination of the image characteristics and matching gray region, the parallax bound and the relative neighbor parallax bound, as well as the continuity and the unique binding constraints. Tests have proved that this technology can eliminate a variety of bad match, and obtain a high correct rate of matching and meet the demand of accuracy.3. The research of camera calibration method. On the basis of analyzing the imaging camera model and its calibration parameters, several typical camera calibration methods are presented. At last, a camera calibration methods based on image distortion correction is proposed and improved to meet the needs of the on-line real-time detection. Experiment shows that the calibration method in this paper is as accurate as Tsai classic two-step camera calibration algorithms. Binocular stereo vision camera is calibrated with the calibration mentioned in this paper. Choosing the three-dimensional objects of Checkerboard grid and using the improved algorithm, this paper finally realizes the rapid and accurate calibration of the parameters inside and outside.4. The main task of this part includes researching the detection principles about symmetrical points of the automobile body and vehicle axle deviation angle, analyzing the relation between vehicle axle deviation angle and the vehicle wheelbase and describing why use vehicle axle deviation angle to instead the vehicle wheelbase.5. The design of detection system. Combined with the performance evaluation objective of visual inspection system, software/hardware program of the automobile body-deviation inspection system based on stereoscopic visual should be rationally designed. According to the actual choice of binocular vision camera model, the structure parameters should be designed for large-size measurement. And the hardware system should be designed, based on the PC platform, by the data of sensor and other subsidiary facilities. Combining the visual theory with the practical inspection tasks, this paper completes the division of the monitoring and control software modules and the design of processes and procedures.6. The research of inspection method about the height difference of symmetrical points of the automobile body and vehicle axle deviation angles based on stereoscopic vision. The test result indicated that the inspection method in this paper can measure the height differences and vehicle axle deviation angles of symmetrical points rapidly and accurately.The main innovativenesses of the study are as follows:1. The testing method of the position deviation of vehicle-body and axle based on stereo vision is putted forword first, and applied and awarded the invention.2. In this paper, the traditional algorithm is analyzed and the improved algorithm is proposed to achieve the extraction of edge, corner and the straight-line accurately. The main assignment of this paper includes that having studied the matching technology in the stereo vision and proposed the stereo matching method based on parallax positioning bound. Tests have proved that this technology can eliminate a variety of bad match, and obtain a high correct rate of matching and meet the demand of accuracy.3. A camera calibration methods based on image distortion correction is proposed and improved to meet the needs of the on-line real-time detection. Using the improved algorithm, this paper finally realizes the rapid and accurate calibration of the parameters inside and outside.4. The testing method of using the vehicle axle deviation angles appraise two axles whether or not parallel is putted forword first, analyzing the relation between vehicle axle deviation angle and the vehicle wheelbase and describing why use vehicle axle deviation angle to instead the vehicle wheelbase.The research and the development work made by this paper is a typical example of application which makes a number of research results in the technology and algorithms. And it has a theoretical significance and the practical application value to raise the stereoscopic vision examination technical level. At the same time, this paper leads the stereoscopic vision into the detection of vehicle body and axial deviation angle, realizes the automation process of vehicle detection. The results of this paper play an positive role in the development of the automobile detection.