Drill Pipe Counting Method Based on Local Dense Optical Flow Estimation

Abstract

To solve the problem of drill pipe counting in coal mines, we propose a drill pipe counting method based on local dense optical flow (LDOF) estimation. Compared to the general tracking method, we provide a new perspective for resolving the problem of drill pipe counting. Taking into account the various working conditions of the drill pipe movement and the complex lighting environment in coal mines, we leverage the local dense optical flow to obtain the movement information of the drill pipe. The model is further established through optical flow to realize the counting of drill pipe. In order to validate the proposed approach, we collect 5 videos of drill pipe movement in coal mines, which cover strong light source and occlusion situations. The experiments demonstrate that our method can realize automatic counting of drill pipe, with an average drill pipe counting accuracy of 97.92% and an average frame per second (FPS) rate of 64.2.