Image Distortion Correction Based On Multiple Free-plane Calibration Methods Based On Particle Swarm Optimization

The inherent properties such as the refractive index of the lens crystal material of the camera,etc.,the unevenness of the smoothness caused by the manufacturing process of the lens surface,and the randomness of the man-made assembly inevitably lead to imaging geometric distortion.The original information of the image is affected,and the real information of the object can't be obtained when the image is used for subsequent processing.Therefore,it is of practical significance to study the distortion correction technique.SF6 circuit breaker is one of the key equipments in the power system.Regularly inspecting the gas pressure and density values is required in rules of the electric power system,but the pressure gauge position is too high or located at the interlayer in substation,which will make it difficult for the staff to get the indication visually.Thus,it is more significant to real-time monitor operating parameters of electrical equipment by combining machine identification method and remote vision technology.In this paper,the image distortion correction method based on particle swarm optimization algorithm and multiple free plane calibration method is studied for image geometric distortion caused by camera lens imaging,and it is applied to industrial image machine identification.This paper analyzes the causes of camera lens distortion,distortion type and nonlinear distortion correction method.The parameters of camera model are calibrated by using multiple free planes calibration method.The inertia weighting factor calculation method based on target value is given in terms of the particle swarm optimization algorithm.The improved particle swarm optimization(PSO)algorithm is used to optimizing the calibration parameters of the camera's mathematical model.Finally,the iterative method and the Otsu algorithm are applied to segmenting the image of the calibrated SF6 circuit breaker pressure dial.The image edge information is extracted and compared from the segmentation result by some methods such as the Sobel operator,the Canny operator and the morphological method.The pointer is located through the straight line and circle extraction method based on the Hough transform,thus the pointer indicator angle is transformed into the pressure indication to achieve the visual inspection in the power system intelligent remote application.The experimental results show that the distortion image can be better recovered by the technology of image distortion correction based on the improved particle swarm optimization algorithm optimizing calibration parameters of multiple free planes calibration method.The result of image preprocessing provides a guarantee for the realization of better segmentation and edge information extraction.The connotation of the telemetry technology is enriched through integrating the visual sensation into "four remote" technology of the power system,which replaces the manual inspection and reduces the visual work intensity of the remote video supervisor.