节点文献
基于人工神经网的激光三维扫描技术研究
Laser 3D scanning technique based on artificial neural network
【摘要】 在激光三维测量技术中 ,用摄像机对物体和物体在平面反射镜中的虚像成像 ,用人工神经网技术实现图像坐标系到世界坐标系的映射 ,无须测定摄像机和激光平面之间的相对位置关系 ,自动修正镜头的几何畸变 ,由单视点序列影像就可获得被测物体表面的三维坐标。结论 :盲区小、方法简单 ,变换关系清晰 ,能获得高质量的测量结果。
【Abstract】 In laser 3d scanning system,a mirror is introduced to reduce blind zone and a CCD video camera is used to take pictures,besides that artificial neural networks are enrolled to map coordinates from screen coordinate system to world coordinate system.During this procedure,lens distortion is calibrated automatically,and without determining the relative positions between the camera and the laser beam the stereo coordinates of 3D surface can be obtained from pictures taken by single camera.Conclusion:With the characters of simple in method,clear relationship between coordinates transformation,quality results can be obtained.
【Key words】 artificial neural networks; laser; 3D measurement; reflector; articulator;
- 【文献出处】 激光杂志 ,Laser Journal , 编辑部邮箱 ,2005年01期
- 【分类号】P234.1
- 【被引频次】2
- 【下载频次】191