Research Of AUV Motion Video Stabilization Method

With the fast progress of the science and technology, the image processing andintegrated electronic circuits technology are developed constantly, the electronic imagestabilization technology is used widely in the military field and civilian field. As the newestimage stabilization method, the electronic image stabilization makes up the disadvantages ofmechanical al image stabilization method and optical image stabilization method veryeffectively. The research of the motion vector estimation algorithm is very important inmethod of the electronic image stabilization, it determines the system’s processing speed andimage stabilization accuracy. To meets the video terminal display system’s performancerequirements of the AUV, this paper has made some improvements on the basis of the typicalgray projection algorithm.Firstly, this paper introduces the basic principle and the range of the applications in thereal life of the three currently used image stabilization methods, analyzes the reasons ofcausing the video image fuzzy, introduces the basic principles of electronic imagestabilization method and the evaluation criteria of the system’s performance.Secondly, investigates the basic principles of translational motion electronic imagestabilization algorithm, and use the MATLAB to program and simulate, analysis theadvantages, disadvantages and the range of applications of them.Thirdly, research the basic principles of rotary motion electronic image stabilizationalgorithm, and use the MATLAB to program and simulation, analysis the effect of imagestabilization processing by the method of minus shadow.Lastly, to meet the video terminal display system’s performance requirements of the AUV,this paper has made some improvements on the basis of the typical gray projection algorithm,first, preprocess the image, use the median filter method to filter the noise of the image, andthen use the method of histogram equalization to enhance the contrast of the image, therebydeveloping the stabilization accuracy of the image. Next, use the fast searching method todevelop the speed of the electronic image stabilization according to the characteristics of thegray projection mutual correlation curve. Then use the method of average filter whichwindow length is adaptive to filtering the estimated video global relative trajectory. Finallyuse MATLAB to program and simulate the improved gray projection algorithm, analysis andevaluate the effect of image stabilization by the method of minus shadow, it proves that theimproved gray projection algorithm is more accurate and faster than the typical gray projection algorithm, it meets the video terminal display system’s performance requirementsof the AUV well.