Pan Hongda

An Image Enhancement Method Based on Gamma Correction

The research of the wheel alignment system based on computer vision is a popular field for automobile safety. For realizing the effective and accurate feature recognition for the chessboard calibration target of the system, the enhancement algorithms based on Gamma correction are indicated. According to the research object, identification requires decreasing the pixel values in low grayscale and increasing the pixel values in high grayscale while keeping the pixels values in the middle grayscale. For this purpose, the novel Gamma correction curve is presented to enhance the target image. Experimental result shows that the enhancement method increases the contrast ratio of the chessboard target image which is benefit for feature extraction, points matching and vision measurement. Its effect and reliability meet the requirements of the detecting system for vehicle wheel alignment based on computer vision.

A Feature Points Matching Method for Calibration Target Images

Feature matching is a crucial research in computer vision. For realizing the effective and accurate feature matching for the feature points in the images of chessboard calibration target used in the wheel alignment system, the popular calculation algorithms are analyzed. The traditional polar constraint and polar equation for solving the problem constructing the relationship between corresponding points are indicated. A novel process which is special for matching calibration target images including sequential consistency constraint, polar constraint, distance constraint, uniqueness constraint are presented. Experimental results show that the matching precision and reliability of the method meet the requirements of the vision detection system for vehicle wheel alignment.

A Novel Corner Point Detector for Calibration Target Images Based on Grayscale Symmetry

The research of feature corner recognition for the calibration target is a crucial area for a vision-based measurement system. For the disability of detecting the corner’s location in a vehicle test system accurately, the theoretical errors based on Harris algorithm are analyzed. Experimental results show that the method can not be used in the vision-measured field as the corner cluster phenomenon. The novel corner detection algorithm based on grayscale symmetry is presented. The Gassian weighted symmetry detector is designed to illustrate the symmetrical feature around the optimal corner. Further Experimental results show that the detector proposed can be used to test the corner locations of the calibration target precisely.

Research on Calibration Theories and Experiments of Vehicle Wheelbase Difference

Considering the disability of calibration theories and methods for the tester of automobile wheelbase difference, mechanical analysis of the vehicle stability is employed to dedicate the negative influences of wheelbase difference. A novel device can realize calibration of the wheelbase difference within the range of 0-40 mm by variable wheelbase mechanism is proposed. Two diagnosis systems which suit the situations that the vehicle is driven in a straight line or a curve are presented. Calibration results of the tester show that maximal detection error of the measured points reaches 0.57 mm which is more than ±0.2 mm in the calibration standard. Experiments prove that the calibration methods proposed can verify the results of wheelbase difference tester fast and accurately.