Research On Key Technologies Of Single Camera Photogrammetry

In recent years,industrial measurement methods combined with computer technology have been paid more and more attention,including industrial photogrammetry technology.The main research object of this paper is the single camera photogrammetry system.Its measurement means are as follows: place the reflective feature mark and benchmark ruler on the surface of the object to be measured,and take photos of the object from different angles with the industrial measurement camera;The three-dimensional data of the marker point cloud is reconstructed according to multiple photos,and the geometric information of the object to be measured is obtained through the three-dimensional coordinates of the point cloud,so as to achieve the purpose of size measurement.This paper focuses on the main key technologies of single camera photogrammetry system.On this basis,it is realized by programming,and a complete single camera photogrammetry system is built to verify the algorithm.The specific work is as follows:(1)This paper studies the sign decoding and positioning technology based on image processing,and designs an improved decoding and positioning algorithm for Schneider ring coded sign,including the preprocessing operation of the original image,the correction and restoration of the sign image,the center positioning of the sign image,and the extraction and decoding of the code value.(2)The three-dimensional reconstruction algorithm of marker point cloud in single camera photogrammetry system is studied,including: establishing the initial model by using the relative orientation of the first two frames of photos,adjusting the threedimensional data of camera internal parameters and point cloud by self checking beam method,solving the initial external parameters of photos by DLT method,non coding marker matching reconstruction based on epipolar constraint,and adjusting the scale of the global model according to the reconstruction of benchmark.(3)Experiments are carried out to verify the coding mark decoding and positioning algorithm and point cloud three-dimensional reconstruction algorithm proposed in this paper.The experimental contents include: the recognition accuracy and efficiency of the decoding algorithm program in this paper are tested under different shooting tilt angles,and compared with the traditional decoding algorithm;According to the threedimensional point cloud reconstruction algorithm in this paper,the reconstruction experiment of the workpiece to be measured is carried out,and the source of measurement error is analyzed while evaluating the accuracy of photogrammetry system.