| Binocular positioning system is one of the important applications of computer stereo vision.Similar to human vision,it has the function of sensing the depth of the surrounding environment through two-dimensional image information,and it has a wide range of applications in the field of computer vision such as visual navigation,motion analysis,and industrial inspection.In the traditional binocular vision positioning system,there are some key technical problems such as difficulty in stereo matching,large amount of calculation,and low efficiency.In order to improve the real-time and robustness of the binocular positioning system and improve the positioning accuracy,a binocular positioning system based on the ROI extraction of the visual image and the detection of SURF feature points is designed.The main research work of this paper is as follows:(1)Describe the distance measurement principle of the binocular positioning system.Based on the requirements of real-time and high-precision research in this project,a binocular positioning system is designed.At the same time,the hardware composition and software flow of the system design are summarized.(2)The selection and design of camera calibration model are carried out.The linear and nonlinear models of the camera are studied and analyzed.The conversion relation between the related coordinates and the coordinate system is clarified.The Zhang Zhengyou calibration method is selected,the calibration procedure is described,and the corresponding calibration experiment and stereo calibration experiment are completed.(3)The principle and technique of segmentation of the region of interest in the image are studied.Since the existing image segmentation algorithms are all proposed for a certain class of problems,they have certain limitations and are difficult to apply to the actual research of the subject.Therefore,according to the practical application background of the subject,a method based on combination of histogram threshold and projection is proposed to solve the key technical problems of ROI segmentation in binocular vision images.Using the gray distribution feature of the ROI to be segmented,adaptive threshold segmentation is performed according to its position in the histogram,and the segmented picture is vertically projected and horizontally projected respectively to obtain the precise pixel coordinates of the ROI in the left and right images.By comparing the size of the left and right ROI regions,ROI segmentation is carried out on the left and right images with the large area as the standard,which lays the foundation for the extraction and matching of the following image feature points.In order to verify the anti-jamming ability of the algorithm,a large number of experiments were carried out under different scenes,different brightness and multi-target images.The experimental results show that the ROI can be segmented out by using the method proposed in this paper,which can meet the actual application scenario.(4)The principle and technical methods of feature extraction and stereo matching based on SURF algorithm are studied.Based on the application conditions and background of this topic,an improved SURF is used to achieve the extraction and matching of sub-pixel corner points in the ROI.Experimental results show that this method greatly improves the matching accuracy and matching efficiency.To sum up,this article is based on the key technology research of the binocular positioning system based on the region of interest(ROI)segmentation.According to the needs of the subject,the binocular positioning system is designed and designed to complete the experimental verification.The software based on OpenCV computer vision library and Visual Studio 2012 is developed.The functions of image acquisition,camera calibration,ROI segmentation and stereo matching have been successfully realized,which meet the requirements of high precision and real time of the subject. |
PDF Full Text Request