Study Of Counting Methods Of The Round Steel Pipes Based On Image Recognition

The paper implements counting round steel pipe automatically through recognizing the image of round steel pipe section.Aim at the conglutination phenomenon in steel pipes, the paper researched the method of the erosion operation based on mathematical morphology so as to separate the pipes. After using the erosion operation can not only separate the pipes at the tenuous area but also eliminate the noise. Using the edge modal detection method of binary image mathematical morphology can detect the edge of steel pipe's inner radius.Based on global geometric symmetry, the original Hough transform(HT)is improved, and a 3 stage approach to extract ellipses or circles is presented. The method significantly reduces the time and space required to evaluate a11 possible parameters and improves the accuracy. The experimental results are given to show the correctness of the method.Hu's moment invariants are a method based region .In the paper, Hu's moment invariants are used to detect the pipe and it achieves high rate.In term of the pipe image that is eroded time after time, the steel pipe's inner circle changed into a point, eight-connected edge tracking algorithm are used to mark and count the round steel pipes.