Study On Key Technologies To Enhance The Performance Of The Laser 3D Body Scanner

The laser 3D body scanner, features non-contact, fast speed, high precision and intelligence, is of important value in application and broad prospects for development. It is therefore widely used in garment industry, medical plastic, and automotive interior design, among others.This paper presents the large amount of research done on the key technologies of a laser 3D human body scanning system prototype, and the enhancement to system performance as a result of the research.The main innovations in the research are as follows:1. An enhancement to the gradient barycenter method, which aims at extracting the laser stripe center, is put forward. This method is capable of handling all types of light band images of a human body and completing the calculations at high speed .It is applicable to the laser 3D human body scanning system.2. A 3D dynamic precision calibration target is designed and manufactured. In whatever form the target moves on the horizontal level, an adequate number of accurate 3D calibration points can be obtained in every direction. Thus, the target can be used in local calibration by inside and outside parameters of the camera, global calibration by multiple 3D sensors, and also combined calibration by multiple 3D sensors and color sensors.3. A camera distortion correction method based on direct linear conversion of the coplanar points is presented. This method takes full consideration of lens distortion on the basis of linear calculation of copular points. With the initial parameter values determined by the camera, the optimal parameter values can be obtained through amendment of the conversion matrix and the distortion coefficient by the least square algorithm. Then, using the dynamic precision calibration target and the distortion correction method presented, calibration is completed on the laser 3D human body scanning system.4. A standard combination is designed to test the performance of a laser 3D human body scanning system. The test results proved the accuracy of the cloud points obtained through scanning of the objects by this system. Finally, a scanning experiment is done on a human model, and the results proved the validity of the theory and the algorithm put forwarded.