Study on Underwater Image Conversion Algorithm Based on Light Refraction Model

Aiming at the problem of image distortion caused by light refraction during underwater imaging, and the conversion error caused by existing image conversion algorithms due to neglecting the secondary refraction of light, an underwater image conversion algorithm based on the light refraction model is proposed in this paper. The algorithm firstly obtains the pixel information of the underwater image, then calculates the corresponding coordinate information of the pixel points in the equivalent air image through the mapping relationship, and finally obtains the equivalent air image through image interpolation. Experimental results shows, compared with the existing image conversion algorithms, the proposed algorithm reduces the average error ofudirection from 2.289 5 to 1.213 3, which is a decrease of 47.01%. The average error ofvdirection is reduced from 3.252 5 to 1.526 3, which is a decrease of 53.07%. At the same time, the mean value of ranging error was reduced from 58.83 mm to 28.88 mm, a decrease of 50.91%。