Field Line Detection Based on Local-precise Extracting and Modified Hough Transform

In Robocup, the field lines provide robots with the most useful information to make self-locating and strategies. For real-time competitions, the detection must be efficient. In this paper, we present a novel method of line detection based on fast edge extraction and modified Hough Transform. In this study, the condition of different objects' distances is considered according to the camera angle. In the first step of edge extraction, an image is roughly scanned. A local-precise scanning is then implemented in regions where there may be edge points. After all feature points are picked out, a modified Hough Transform is used to extract lines. The result demonstrates that combining with real-time conditions, such as the angle of the camera and the distances of the objects on court, plays an important role in object detection.