挖掘机器人视觉控制系统研究
Exploring Visual Control System of Excavator Robot

对Labview建立的挖掘机器人视觉控制系统在目标物发生旋转、比例发生缩放以及光照条件发生改变时无法始终对目标物进行识别与定位的问题，提出了SIFT图像匹配算法。将其特征描述子通过高斯核函数映射到了更高维的特征空间上去，然后在该空间上对其数据进行降维等改进处理。通过Labview中的Matlab script和Read From Spreadsheet File VI节点调用改进后的SIFT M文件，实现了Labview与Matlab的无缝联接。实验表明：Labview 与Matlab混合编程的挖掘机器人视觉控制系统解决了原先系统所存在的问题，其功能更加完善和强大。
To solve the problem that its visual control system cannot identify and locate the targets all the time when they perform rotation and scaling or when the lighting condition changes, we proposed the SIFT image matching algorithm and mapped its feature descriptors to a high-dimensional feature space through using the Gaussian kernel function, and then enhanced the targets in the feature space, for example, data dimension reduction. The enhanced SIFT image files were invoked into Labview through the Matlab script and read from spreadsheet file VI nodes, thus realizing the seamless connection of Labview with Matlab. The experimental results show that the Labview and Matlab hybrid programming of the excavator robot's visual control system can solve the problems existing in the old system and is more perfect and powerful