The Design And Implementation Of Fisheye Image Correction Based On Specific Angle

Recently,high-definition video camera has atractted the attention of the market.The demands of the users have changed from see clearly to get full scale of the secne.In order to achieve this requirement,traditional video monitoring approach installed several cameras in the same area and using the scan method to get full scale of image.This method partially sovled the problem but requires significant human work and devies,which resulting in a high cost.Fisheye lens have ultra wide angle panoramic vision and is widely used in construction.It is one of the most simple and efficient method for lagre scale imging.In this thesis,with the study of the fisheye image correction algorithm,a simple and effective,fast and practical fisheye image correction method is designed.The method achieved a 360 degree correction of the fisheye panoramic image.The main contributions of this paper includes:1.The correction algorithms of fisheye image under horizontal angle of view are studied: the latitude and longitude mapping method and the equidistant projection method.Two algorithms are implemented by programming,and the correction effect of the algorithm is verified.From the results,both algorithms can achieve fisheye image correction.But to generate a panoramic image with a horizontal direction of 360 degree,at least two fisheye images are required to splice.The fisheye image captured in the vertical direction contains all the information of the target area.In order to reduce the workload and improve the efficiency of generating panoramic images,the panoramic correction of fisheye images with specific angle is selected as the main research direction.2.The methods of fisheye image effective circular area are studied.After comparing the advantages and disadvantages of each method,the scan line approximation method is selected to obtain the fisheye image effective circular area with fewer calculations and better extraction effect.The initial acquired fisheye image is corrected to correct the contours of the fisheye image as a standard circle.3.Two panorama expansion algorithms based on specific angle fisheye images are implemented and compared: the fast expansion algorithm and the 30 degree rotation expansion algorithm.In order to satisfy the real-time requirement of engineering,this paper proposes a horizontal mapping method based on spherical perspective to realize the panoramic correction of the fisheye image.The basic idea of the algorithm is that the pixel on the fisheye image is projected onto the spherical model and then the perspective image is projected onto the VLP to generate a panoramic image.The effect of panoramic expansion of fisheye image is ensured and the time of panoramic image generation is greatly reduced.4.Based on the spherical perspective algorithm,the panoramic expansion scheme of the fisheye image is designed and implemented.The feasibility of the proposed scheme is verified by several experiments.In this paper,the panorama development of the fisheye images of multiple scenes is carried out.The time costs is recorded and the three-dimensional display effects are compared.The performance of the algorithm is verified and the feasibility of the scheme is proved.