Measurement Method of Underwater Target Based on Binocular Vision

摘要

Traditional measurement methods are obtained by calculating the global disparity map. The accuracy of these methods is low and the speed is slow. Under the condition of uneven illumination, low contrast and obvious noise, they can't meet the measurement requirements well. An underwater image enhancement method based on Contrast Limited Adaptive Histogram Equalization and image weighting fusion is proposed. For the measurement of underwater target, only two matching points are solved in the algorithm design, which improves the accuracy and speed of underwater ranging. The idea of coarse-to-fine matching is adopted, that is, the SAD algorithm is used in matching area to complete the coarse matching such that the matching windows can be determined firstly. Then the stereo matching algorithm based on NCC and Census fusion is used to decide the final matching point. The algorithm is insensitive to light and can extract edges and comers better. Finally, three-dimensional reconstruction is carried out to restore the three-dimensional coordinates of the endpoints to be measured, and the real size can be obtained by calculating the Euclidean distance between the two points. The experimental results show that the algorithm runs fast and robustly without calculating the disparity of the whole image, and the results meet the accuracy requirements.