Efficient and Accurate Coded Target Decoding for 3D Reconstruction of Soil Specimens in Triaxial Test

Coded targets have been widely used for solving the corresponding problem in photogrammetry for high-accuracy three-dimensional measurements. Accurate and efficient recognition and identification of coded targets are of great importance in coded target-based photogrammetry. In this paper, an efficient and accurate method for coded target decoding was developed. In this method, blob analysis was performed to recognize the coded targets. Then, image processing, the RANSAC algorithm, and the interpolation technique were applied respectively to decode the coded targets, identify falsely identified coded targets, and recover missing coded targets. Interpolation was also performed on the membrane, which can significantly increase the density of the points on the membrane and produce more representative 3D results for the soils specimen. This method was implemented into a MATLAB program and all computation was done automatically by the computer program. This method takes advantage of the prior knowledge of the geometric arrangement of the coded targets. The effectiveness and accuracy of the proposed method are validated by implementing it into three-dimensional reconstruction of soil specimens during triaxial testing in geotechnical engineering. Experimental validation results indicate that the proposed method can achieve accurate and efficient coded target recognition and identification results.