Robot fingertip tracking process analysis and simulation

The technology of fingertip tracking is widely used in robot control field, the fingertip of robot can be accurately controlled. But the fingertip tracking is very complicated, there is much uncertainty for the speed and direction during the movement of fingertip, so the fingertip tracking technology becomes one key point in the robot control field. In the early, the feedback tracking algorithm was widely used, but there is big deviation between the tracking result and the actual movement process. With the development of visual image processing, using the image processing method to do the fingertip tracking becomes popular. To avoid the low accuracy defect that comes from the traditional algorithm doesn't consider the light intensity during the fingertip tracking process, this article provides one new fingertip tracking algorithm based on image processing. For the collected image of robot fingertip tracking, first perform the initialization to get vector of the high frequency vector and low frequency vector, then perform the wavelet processing for the collected vector to improve the image quality. Then normalize the pixels of fingertip tracking image to provide accurate data for the finger tracking. At last, use linear analysis method to process the data from normalized data. The experiment result shows that the tracking accuracy can be effectively improved by using the improved algorithm on fingertip tracking.