Research On 3D Reconstruction Technique Based On Microarray Images

Three dimensional reconstruction is a hot issue in the field of computer vision.The common methods of 3D information acquisition are structured light,laser scanning,ultrasonic imaging and other active methods.And the passive ways are recovering shape from shading and texture,and recovering structure from motion.Compared with other 3D information acquisition methods,3D reconstruction based on image has many advantages such as high efficiency and wide application.Large scene modeling,such as three-dimensional topography,military terrain,urban streetscape,ancient architectural model,etc.can be quickly generated by multiple two-dimensional images.As more and more images are acquired by mobile devices,the status of mobile phones is becoming more and more prominent.Therefore,it is of great practical significance to study the 3D reconstruction based on microarray camera.In addition,it has the characteristics of short baseline,similar field of view and the demand of high precision of matching,so it has important scientific value.The main work and innovation of this paper are as follows:(1)The calibration of microarray camera system.Camera calibration includes active calibration and self calibration.The main difference is whether the camera motion information is needed.In this paper,a new five point calibration method is proposed based on the Zhang Zhengyou plane calibration method,which not only improves the accuracy of parameters,but also simplifies the calibration process.(2)The extraction and matching of corresponding points.The feature points can be extracted by SIFT feature points algorithm and Harris corner points algorithm.The scene of microarray images is quite similar and the image resolution is 1600*1200.It can be converted to the tracking problem of video image,using Shi-Tomasi algorithm to extract the corner points.The algorithm can extract more feature points in a short time and the KLT tracking algorithm can increase the matching efficiency.(3)3D reconstruction based on short baseline.In this paper,we study the matching of two images and the matching of 3 images and more.Combined with the advantages of multi-lens camera,the matching accuracy can be greatly improved by using the common matching points of 4?6 images.In the case of only 5?6mm of the baseline length,the solution of 3D point coordinates can be accurate to mm level.Through the research on the 3D reconstruction technology based on the microarray image,we hope that we can do some research on the mobile phone of multi-lens in order to provide users with a richer experience.