| Usually, traditional aerial surveying can be taken only in the daytime, its post dataprocessing is not automatic and needs manual manipulation. Since the1980s, airborne lasermapping system has appeared in Europe, United States and Japan. It can be constructed easilyand has much superiority compared with the traditional aerial surveying such as low cost,faster terrain scanning, environmental adaptability. Laser mapping system can be used togenerate more accurate digital terrain model and digital elevation model. It has been widelyapplied in the route exploration, coastline surveying, the forest resources distributionobserving, digital city construction, location based service, and many other fields.This paper firstly describes the construction of an airborne laser system, the theory ofpoint cloud data processing and the software implementation. It includes data communicationlink protocol, different sensor data synchronization, the extended Kalman filter designed forthe attitude and position, three-dimensional point clouds processing procedure, Delaunaytriangulation algorithm, OpenGL three-dimensional display and selection feedback, etc.Because the airborne laser system involves numerous sensors, software processingbecomes complex. The acquired inaccurate data along the X, Y, Z axis would result insurveyed terrain sloped and periodically distorted, it deteriorates the performance of thesystem. This article analyzes how the various hardware and software processing influence the3D point clouds accuracy. Based on this point, the error model is then established which canbe used to direct the practical experiments. Finally, some methods are proposed for systemerror calibration and elimination (e. g. the attitude angle calibration, placement anglerecovering, adjustment processing, etc.). The total errors include random error and systemerror, by the proposed methods in this paper, most of the system errors can be eliminated andthe accuracy of point clouds is then improved to a great extent. |
PDF Full Text Request