A Kind Of Image Fast Detecting Algorithm For Moving Target

This research is to solve the problem that moving target can't be detected correctly under industrial environment and the problem of speed in image detecting. For example, how to detect the fast moving billet in the heating stove in the steel rolling factory.Based on the theory of statistic image recognition , in this research we put forward a kind of fast detecting algorithm for moving target. Firstly we can detect the simple target by the gray value feature; while we do some research on the feature of gray apex and valley. By this algorithm we can not only detect the simple target fast, but also detect the complex target fast. Then we go on perfecting the algorithm. With gray value and gray apex and valley , we construct the two dimensions gray feature space , and then we can detect the simple or complex object more correctly.With the simple gray features ,this kind of algorithm is so simple and fast that it can satisfy the requirement of fast property of dynamic target detecting. Furthermore , we design the recognizing scheme through known sampling images ,according to the statistical method. Then we can guarantee in theory that the algorithm can recognize the target correctly or the recognition error is minimum to some mathematic degree. So it can also satisfy the requirement of correct property of detection.During this research, in the lab we build up the experimental system of image detecting which simulate the condition of the practical field. We validate the fast property and the correct property of this algorithm by a series of experiments on this system. We also implement the function of detecting and controlling through this system. The result shows that this algorithm has the property of more fast and more correct, so that it could be utilized to measuring object in the industry, meeting the need of dynamic measurement.