Research On Automatic Correction Approach Based On An Single Wide Angle Distortion Image

Compared with the ordinary lens,a wide-angle lens can usually get more perspective scenery information and can capture more information in the image,so it is widely used in road traffic,video detection and medical equipment etc.However,the wide-angle lens inevitably brings nonlinear distortion due to its own manufacturing process and other defects,and radial distortion is the most important distortion and plays a key role in nonlinear distortion.Radial distortion is defined as the deviation of the coordinate position introduced by a pixel in the radial direction so that the image captured by the lens is distorted.The introduction of distortion is extremely unfavorable to the visual effect of the image and the subsequent image processing algorithms such as tracing,recognition and three-dimensional reconstruction.Under this background,this paper proposes a novel automatic correction algorithm based on a single wide-angle distortion image and obtains a good correction effect the image radial distortion by analyzing the current status of the image radial distortion correction algorithm at home and abroad.The algorithm reduces the time complexity and improves the stability and accuracy of the correction compared with other papers.The major research contents of the paper are as follows:(1)In the part of arc extraction from edge contours,this paper presents a distortion line extraction algorithm based on the inherent nature of equal-arc.By using the proper edge detection and contour extraction algorithm for the distorted image,the distorted lines satisfying the intrinsic characteristics are detected in the edge contour by using the inherent characteristics of the equal-arc,and finally the distortion line extraction algorithm is realized.The algorithm can extract distorted lines from edge contour image fast and stably.(2)In the part of distortion parameter estimation,this paper presents a distortion parameter estimation algorithm based on double voting mechanism.By analyzing and researching the set of distorted lines,the length of distorted lines,the high quality and the low quality distorted lines,the distribution of distorted lines and the accuracy of distorted centers are all the factors that influence the distortion parameter estimation.In this paper,we use the constant feature of the distortion center and the high-quality arc to vote and select the optimal interval with the largest constant feature to find the optimal distortion parameter and the corresponding distortion center..(3)In the part of automatic correction of distorted images,this paper presents a robust algorithm for automatic correction of distortion image.The paper adopts the appropriate method of edge detection and contour extraction,and proposes the algorithm of the distortion line extraction and the distortion parameter estimation,and adopts the correct algorithm of the reverse correction of the distortion image.The comparison experiments show that the proposed method of automatic image distortion correction is fast,stable and accurate.