| The object three dimensional outline is containing the object shape information, the measurement of shape and scale is one of basic methods which are used to rebuild the world by mankind. In recent years, structured light 3D measurement has attracted a great deal of attention in industrial environment because of its following characters:high precision, easy to abstract the information from laser stripe image and strong real-time abilitym, etc. This technique is too important to be replaced by others in the field of 3D auto-measurement and object recognition.Basing on large investigation of the three dimensional reconstruction method home and abroad at present, the paper adopts the light-section method and makes object revolving to reconstruct the object, and succeedes to reconstruct the locomotive wheel to achieve the goal of comprehensive detection of the wheel surface around.The main part of the reconstruction system including camera, laser, revolving platform and the computer.The laser sends out a bunch of light to the surface of the object, take pictures using the camera. Because the laser light strip has the object surface outline information, the object outline original information can be deoxidized by gathering the picture and inputting them in the computer to process. If let the object revolve around and gather a laser picture every other certain angle, then may display the entire object three dimensional outline drawing.The three dimensional reconstruction algorithm of object is designed in this paper. It includes image processing, system calibrating and three dimensional reconstruction display. and image processing part mainly including the image preprocessing, image cutting and image thinning. The laser light strip thinning algorithm in the image-preprocessing is the one causation of infection of the precise of the object reconstruction, Therefore it is the key part of the algorithm design. The paper compares the laser light strip geometric center extraction process method and the skeleton method, and adopts the skeleton method whose thinning precision is much higher. During the whole designing processing, the system calibration is not only the key part also the most difficult part of the paper, the precise of the calibration directly affects the precise of the reconstruction. The paper has investigated the advantage and disadvantage of several camera calibration methods home and abroad, and adopts the Zhang Zhengyou calibration method, which is realized quite simple and the precise of the calibration is quite high, and have used a Matlab calibration toolbox which is based on Zhang Zhengyou calibration method to calibrate the camera. Based on this, then the object image coordinate data can be transformed to the space coordinate data, and preprocessed in the Matlab, after the sorting, the curve can be constructed. And using the simple and feasible triangular net grid algorithm realize to the construction of the surface. Finally, the three dimensional model of the object has been displayed in the Matlab and the OpenGl and the result of the 3D reconstruction has been analysised. |