Research on Cotton Flow Velocity Measurement Based on SIFT Feature Matching

In the process of eliminating foreign fibers by foreign fiber machine, estimating the cotton flow velocity can help to improve the accuracy of foreign fiber machine. Linear CCD cameras at different positions are used to collect cotton flow image synchronously. The overlapping parts of two images at the same time are matched through image processing algorithm. A cotton flow velocity measurement method based on SIFT feature matching algorithm is proposed. The advantage of scale invariant feature transformation (SIFT) is introduced into cotton flow feature extraction, which solves the problem that it is difficult to extract cotton flow features. Finally, the speed is calculated according to the actual distance and pixel difference between the two cameras. An experimental device is designed to verify the effectiveness of the method. The cotton flow velocity measurement method based on SIFT feature matching algorithm measures the dynamic cotton flow velocity, and the average error is 5.29%.