Three-dimensional Measurement Of Vehicle Identification Number Based On Binocular Stereo Vision

Engraving of vehicle identification number is a key process in the automobile assembly process,and its engraving quality has an important influence on subsequent processes.According to the requirements of the national standard,the height of the letters and numbers of the engraved vehicle identification number should be greater than or equal to 7.0mm,and the depth should be greater than or equal to 0.3mm.The quality evaluation of traditional vehicle identification number engraving mainly adopts manual measurement or contact measurement,which has low efficiency,poor accuracy and unstable measurement data.In this paper,a low-cost binocular stereo vision scheme is used for three-dimensional measurement of vehicle identification number,which improves the traditional measurement method and has important research significance.In order to establish the mapping relationship between image and three-dimensional space and obtain the feature points of length and width measurement,the camera calibration and image processing work are carried out in this paper.Firstly,the binocular camera is used to collect the vehicle identification number image,and the internal and external parameters of the binocular system are obtained by Zhang Zhengyou’s calibration algorithm.Then the Bouguet algorithm is used to perform stereo correction to make the two camera imaging planes parallel.Due to the influence of environment,there is a lot of noise in the image.In this paper,Gaussian filtering algorithm is used to denoise the image.Considering that there will be different degrees of tilt in the collected images in the real scene,this paper adopts the combination of perspective transformation and plane inclination detection to complete the tilt correction of the image.After the calibration,this paper uses Canny edge detection algorithm to extract the edge information of each character of the vehicle identification number,and uses horizontal and vertical projection segmentation method to segment the characters and determine the measuring points of each character.In order to effectively obtain the engraving depth information of the frame number,a stereo matching algorithm based on multi-feature fusion and guided filtering is proposed in this paper.The algorithm integrates the improved Census transform,pixel color information and gradient information in the cost calculation stage.In the aggregation stage,a guided filtering algorithm with edge preserving property is adopted.The comparison experiments show that this algorithm has good anti-interference ability and effectively improves the matching accuracy.After completing the previous theoretical research,this paper applies the binocular vision scheme to the three-dimensional measurement of the vehicle identification number.Based on the triangulation principle,the length,width and depth information of each character of the frame number are obtained,and the measurement error is calculated.The experimental results show that the measurement error of the binocular stereo vision scheme adopted in this paper is within an acceptable range,and can meet the requirements of the three-dimensional measurement of the vehicle identification number.