Multi-camera Stitching Technique For Mobile Robot

Video mosaic is an extension of technology based on image mosaic,also has been a hot research derection of computer vision,image processing and computer graphics,has been widely used in military,security monitoring,medical operation,GIS system,traffic analysis.At present,there are many methods getting panoramic image,such as Japan Ricoh developed THETA,Google map.Panoramic image can also be implemented by some professional equipment,but the equipment is expensive, limited rotation angle and poor flexibility.First of all,this paper starts from the image mosaic to introduces the basic process of image mosaic and the main steps in detail. The method focuses on image registration and image fusion.This paper focuses on is the problem of real-time multi camera mosaic, image mosaic from the transition to real-time video stitching and introduces some problems in video mosaics, such as selecting the coordinate system, projection model, multi images mosaic synthesis.Secondly,for the multi camera video image sequences are collected at the same time in different view points,this paper studies two kinds of stitching technique based on the feature points.This paper introduces the Harris algorithm and Sift algorithm. The robustness of SIFT algorithm is high and real-time can not be guaranteed.Harrisalgorithm’s registrationcomplexity is low,but it is difficult to solve the problem of poor image quality due to the wide-angle lens distortion,matching performance is also poor.Although the Surf algorithm is fast,but it can’t guarantee the quality.For the same scene.The paper compares the real-time performance of the three algorithms to video mosaic for obtaining the best algorithm.Further improving the algorithm to achieve better real-time performance.Finally,the camera mosaic is used in mobile robot to test algorithm’s real-time.