Research On 3D Measuring Instrument Based On Computer Vision

With the development of modern computer technology,optical image technology and large scale integrated circuit,computer vision technology has been widely applied to industrial manufacture,agricultural production,man-machine interaction and satellite remote sensing and many other fields.As a consequence,non-contact 3D measuring technology has also been developed correspondingly.Due to the advantages of high resolution,no contact and high-speed data acquisition,non-contact 3D measuring technology has gained more and more attention and thus has been widely used.On the basis of computer vision technology,a set of non-contact 3D measuring apparatus was constructed by the author.The measuring apparatus is composed of hardware devices such as analog camera,stepper motors,grating scales,and FPGA-based control modules,image processing modules.High-speed image acquisition and image processing are all completed in FPGA.The use of FPGA improves the real-time performance of the system,meanwhile,it also reduces the complexity of the system.The parallelism of FPGA ensures the real-time processing of the image.The paper first analyzes the working principle of the three-dimensional measuring instrument and designs its hardware structure.Afterwards,different video formats are converted in the FPGA.In addition,this paper,according to the requirement of image processing,selects Gaussian smoothing as image preprocessing and chooses the Canny edge detection algorithm as core processing algorithm.Based on the characteristics of FPGA,this paper optimizes the Canny algorithm with several methods,including the replacement of convolution with calculated template,appropriate approximate transformation and the full utilization of the parallelism of FPGA.In this way,the original good detection effect of the Canny algorithm,the accuracy of edge locating and the advantage of less response on the same side are maintained,the computation speed of the edge extraction is increased while the calculation and operation time are reduced and the real-time performance of the system is ensured.Through experimental verification,the optimized Canny algorithm has obtained satisfactory processing results and completed the real-time processing of the image.Finally,the edge of the extracted image is measured,and the scale coordinates of the edge are accurately recorded in the FPGA to complete the measurement of its size.