Research On Underwater Camera Information Recognition On Technology Of Manned Submersible

The deep-sea environment contains abundant biological and mineral resources.The camera technology is an effective means to study the diversity of deep-sea biological species and the distribution of deep-sea minerals.At present,marine research institutions have installed cameras in some sea areas to achieve dynamic detection of the ocean.However,marine researchers need to invest a lot of time and energy into long-season camera information to extract target information from the camera.Aiming at the problem of how to efficiently mine underwater information,this paper proposes a research on underwater camera information recognition technology for manned submersibles.This technology realizes manned by character extraction and feature classification of specific objects in deep sea camera information.The mining of submarine underwater camera information,the specific research work is as follows:(1)Aiming at the difficulty of extracting underwater camera information of manned submersible,a method of underwater camera information recognition of manned submersible is proposed.A software of underwater camera information recognition system of manned submersible is designed based on underwater camera information recognition technology,which applied computer vision technology and deep learning in the field of manned submersible.(2)According to the characteristics of deep sea image,three different deep learning recognition algorithms are selected,and the principle and structure of the three recognition algorithms are studied and analyzed in detail.Through a large number of experimental data,three kinds of deep learning recognition algorithms are compared for single and intensive models of underwater camera information in the recognition speed and accuracy.The experimental results show that the recognition accuracy and speed of different recognition algorithms are different for the same underwater image information of the manned submersible.Finally,the R-FCN algorithm with the best comprehensive performance is identified as the software recognition algorithm.(3)The software function scheme of the underwater camera information recognition system of manned submersible is designed,and the software functions such as image recognition,data storage,history query and artificial inversion are realized.The overall communication scheme of the system is determined,and communication with other airborne equipment in the manned submersible realizes the correlation between the identified biological information and the geographic information and the sea information,and finally realizes the multi-information fusion function of the manned submersible.