Research On Interior Close Quarter Three-dimensional Laser Scanning Technique

Three-dimensional laser scanning technology has always been one of the research hotspots in the field of machine vision,which can obtain the three-dimensional data of the surface quickly,and realize the reconstruction of three-dimensional scene by collecting the surface information of the measured object.This paper researches the three-dimensional laser scanning technology based on a line laser and a single camera,including the hardware platform design,image processing,the calibration and point cloud data acquisition,the processing of point cloud and so on.The paper uses the line laser to scan the object and capture video stream data by camera at the same time.And then,according to sub-pixel location of line laser stripe and the principle of triangulation to gain point cloud data of object surface.In the end,point cloud data is filtered,matched and reconstructed.The main research work of this dissertation is summarized as follows.1.The hardware platform architecture of laser scanning is accomplished.Based on measuring the range of the points,the paper analyzes and discusses line laser scanning platform.Then,the core devices of the platform are chosen based on the analytical results,such as camera,laser and steering gear,in order to establish the construction of three-dimensional laser scanning platform.2.The recognition and extraction of laser stripe.According to the characteristics of the laser stripe in the scene which is more obvious than other objects in the scene,the paper extracts the red channel in the image for the research.And then,the design realizes rough extraction of the laser stripe through auto-adaptive threshold segmentation and use improved center of gravity to achieve accurate extraction.The algorithm has both faster speed and higher calculation accuracy.3.The acquisition point cloud data.The camera imaging model is analyzed in order to correct distortion of the image by using the method of flat calibration algorithm.Combining triangulation principle with the speed of scanning sampling in three-dimensional laser scanning system,the paper accomplishes measurement of distance and the acquisition of three-dimensional point cloud data of the measured object.4.The processing of point cloud data.This paper does some research on point cloud filtering characteristics and use statistical model methods to remove the noise of the original point cloud data.ICP(Iterative Closest Point)registering algorithm is used to improve the density of point cloud.At last,The three-dimensional model reconstruction of point cloud iscompleted by greedy triangulation algorithm.