The Reaserch Of Underwater Camera Calibration And Measurement Method

With the development of economy and the increase of population, the demand of various kinds of land resources is becoming larger and larger. However, the majority of the earth's oceans contain a wealth of resources that is not developed, and now people are gradually increasing the exploration of the ocean. Because of the dangerous environment of deep sea to humans, some detecting work is done by robots instead of humans. So the images and video captured by the underwater robot are the main sources of exploring the ocean. However, due to the complexity of the marine environment, the quality of underwater photography will be seriously reduced, and the 3D reconstruction and calibration technology for obtaining underwater information is also affected by the water environment.The influence of the refraction phenomenon on the imaging model of the air is introduced in detail, and the physical processes and the mathematical description of the refraction is analyzed too. Then, the nonlinear model is established according to the relationship between the 3D points and the image coordinates in the underwater imaging system. By analyzing the traditional Tsai calibration method and the plate calibration method, the nonlinear model parameters are obtained by adding the nonlinear model to the calibration algorithm. In this dissertation, the accuracy of the calibration algorithm and nonlinear model are verified by the measurement experiments. Then by analyzing the influence of the underwater imaging model on the polar line constraint in the binocular vision system, the dissertation comes to a conclusion that the polar line of the underwater binocular vision measurement system is no longer a straight line. Then, a method is set up to correct the images taken by the underwater binocular vision measurement system, so as to restore the image in the air, and then carry on the binocular vision measurement.The dissertation analyzes underwater imaging system and established the nonlinear model, the model is an accurate imaging model, which does not exist in the process of approximation. And the model is calibrated through the improved calibration algorithm. The accuracy of the model is verified by the experiment. Based on the underwater imaging model, a method of underwater 3D measurement based on binocular vision is established, and the method of underwater image correction provides an important basis for other applications based on binocular vision.