Target Pose Measurement Study Based On Monocular Vision Of Linear Structure Light

Vision measurement is widely used in various fields of inspection, industrial automation production line, visual navigation, 3D four-wheel alignment. Structured light monocular Visual measurement is a typical method of machine vision measurement.High-precision machine vision measurement can be obtained through high-precision camera calibration. Structured light measuring precision is influenced by many parameters such as image feature point detection accuracy, calibration method, hardware devices and so on.Theoretical analysis and experimental validation are used to study the method of precision improvement when the objects rotation angle are measured by structured light sensor, among which feature extraction, structural parameters optimization and data processing are three starting points, graphic target and automobile wheel are experimental measurement objects. The main contents are as follows:(1) The research of structured light measurement method research. This paper mainly expounds the composition and measurement principle of structured light measurement system of the target position, combined theoretical knowledge of monocular vision with structured light, establish the structured light vision measurement model, and analyze the influencing accuracy factors of structure light measurement, which lay a foundation for the structural optimization of simulation analysis and subsequent experimental measurement.(2) Study on the method of image feature point processing. The image preprocessing methods, feature extraction method and the light target corner extraction method are described. Then, according to the characteristics of structured light image, a method to extract image features is proposed.The initial model is setted around the light strip center line, the Gauss model as a template, the initial model of light energy of the optimal solution is obtained, the stripe center extraction method, the geometry information of the center of light stripe with light gray to calculate.The method can effectively suppress the influence of noise, and has strong the light break repair capacity, the extracted light stripe center has single connectivity, sub-pixel precision. The result of anti noise and extraction is good by simulation and experiment.(3) Calibration and optimization of structure parameters of the structure light measurement system. Methods of camera calibration and the light plane calibration are introduced in the paper.Through a series of experiments,a suitable camera internal parameters calibration method is selected. The influence parameter of system measurement precision(internal parameters of cameras feature extraction accuracy, the structural parameters of the system) is analyzed, and structure parameters is optimizated by simulation.(4) Experimental verification.Experimental verification of the research content is conducted on the platform built by three-coordinate measuring machine. The results show that, the maximum displacement deviation and angle deviation of the proposed method are less than 0.08 mm and 0.1 degree respectively. The method has high precision that it can meet the requirements of industrial test.