Research Of Point Cloud Data Acquistion Technology Based On Asodvs

Three-dimensional laser scanning technology is also called real scene replication technology. The technology is widely concerned, because it can measure complex surface effectively and accurately without touch. Traditional three-dimensional laser scanner with high resolution can obtain point cloud data quickly from object surface of large area by high speed laser scanning. It provides a new method for three-dimensional information acquisition. But the application and development of the device is limited due to its defects of high price, limited scan range in one scan and complex data splicing.To overcome the problems above, this thesis designed an Active Stereo Omni-directional Vision Sensor(ASODVS) to acquire point cloud data of object surface. The ASODVS is integrated mobile planar laser generator and ODVS with the characteristics of single view point and 360 degrees Omni-directional imaging. The device can obtain position information and color information of scene point cloud in all directions at the same time. It improves the efficiency of point cloud data acquisition. So this thesis does an in-depth research on ASODVS. The main work of this thesis is as follows:1. Designing an ASODVS system, basing on imaging principle of single view point ODVS. It can realize the automation of acquiring position information and color information of scene point cloud in all directions through controlling the scanning parameter by system software. The point cloud data obtained in scanning is orderly stored by azimuth and height.2. Studying laser stripe center extracting algorithm, which is an important technology of point cloud data acquisition based on ASODVS. Many different laser stripe center extracting algorithms for omni-directional slicing images are been used to compare with each other by experiment and choosing the most suitable algorithm of the system.3. Analyzing the system ranging accuracy and studying the sensitive to the degree of systematic errors of various factors based on the principle of measurement of ASODVS. Presenting the measures to improve system ranging accuracy for the results.The contribution and innovation of this thesis are mainly reflected as follows: firstly, reducing human intervention in the process of point cloud data acquisition and increasing the degree of automation; secondly, acquiring all the point cloud data in one scan utilizing the characteristic of wide field of view of ASODVS and avoiding the multi-angle sets of point cloud data matching; thirdly, acquiring the position information and color information of point cloud both in the same coordinate, avoiding the situation that combining two kinds of data in different coordinates.