A Feature Extraction Algorithm Based on Structural Light in Sea-air Imaging

Abstract

The distortion phenomena in sea-air imaging mainly include geometric distortion, color deviation and low contrast. In order to obtain the target information accurately, it is of great significance to study the distortion image correction. More feature information can be obtained by using grid structured light. However, the random fluctuation of water surface and the complexity of underwater imaging environment, the traditional corner detection algorithm can not effectively obtain the correct corner information. In this paper, a fast corner detection algorithm based on minimum kernel value (MRSFAST) is proposed. The experimental results show that compared with the traditional detection algorithm, the algorithm can extract the feature point information of the image in real time, which is of great significance to further realize the distortion image correction.