Visual Measurement Method of Image Stitching Based on Homography Matrix and Image Quality Evaluation

XIN LI, LIMING WU, GUITANG WANG, YUJUN CHEN, JIANHAI LIN, KUAI RONG

Abstract


Aiming at the problem of low precision of vision measurement technology for large size complex parts, a method that visual measurement method of image stitching based on homography matrix and image quality evaluation is proposed. Firstly, the normalization model between the world coordinate system to image coordinates is established to extract the simplest expression matrix H. Then the camera is calibrated to determine the pixel equivalent P, the field of view and obtain the focal length f. Images of component are obtained, and registration between the stitching images is obtained by the simplest expression matrix H to complete the stitching in many images. Result in this method is compared with the results of SURF and backward mapping model normalization, and the SSIM value of stitching images is calculated to evaluate the quality of stitching images, and the quality of stitching which meets the subjective feeling has been obtained, which meets the requirements of large size parts inspection.

Keywords


Vision measurement, image normalization model, homography matrix, camera calibration, image mosaic, mass analysis


DOI
10.12783/dtcse/aiea2017/14913

Refbacks

  • There are currently no refbacks.