|
|
Detection and extraction of compaction edge of roller steel wheel |
School of Construction Machinery, Chang′an University, Xi′an, Shaanxi 710064, China |
|
|
Abstract For the unique unstructured working scene and relatively fixed working track of the roller, the visual processing algorithm was adopted to extract the rolling trace of the steel wheel. The RGB image of rolling area was grayed, and the random noise was eliminated by preprocessing algorithm. By comparing Roberts, Sobel, Prewitt, Canny, Kirsch and other edge detection operators, the operator with the best detection effect was selected. Hough transform was used to extract the rolling trace line segments. The results show that the edge detection effects of Canny and Kirsch operators are more clear. After Hough transform, the straight line error extracted by Kirsch operator is the smallest. The method combined with Kirsch operator and Hough transform can extract the edge information of steel wheel accurately and can ensure the accuracy and realtime, which provides important visual information for the automatic driving of roller.
|
|
|
|
|
|
|
|