摘要
A method for maneuvering target tracking based on in- ductive factor of posture information is proposed. A distinguished charactedstic of video frequency tracking is that it can capture the target posture changes from its picture easily, and the posture change means the motive model of the target will change. This information is very important to predict the trace of maneuvering target. Based on this idea, the quantified values of the target pos- ture change are obtained using Hough algorithm, this key values are defined as inductive factor of posture information, and then, the multiple grey trace predict models are established and the degrees of fuzzy subordinate values for every model are calculated with the inductive factor, the maneuvering extent values are determined by a new analysis method of stochastic differential equations for each model used to modify the degree of fuzzy subordinate values, these constitute the weighted values for every grey predict collec- tion. Finally, the synthesis predicting weighted result is obtained. The experimental results show that the new method is superior to the conventional algorithm.
A method for maneuvering target tracking based on in- ductive factor of posture information is proposed. A distinguished charactedstic of video frequency tracking is that it can capture the target posture changes from its picture easily, and the posture change means the motive model of the target will change. This information is very important to predict the trace of maneuvering target. Based on this idea, the quantified values of the target pos- ture change are obtained using Hough algorithm, this key values are defined as inductive factor of posture information, and then, the multiple grey trace predict models are established and the degrees of fuzzy subordinate values for every model are calculated with the inductive factor, the maneuvering extent values are determined by a new analysis method of stochastic differential equations for each model used to modify the degree of fuzzy subordinate values, these constitute the weighted values for every grey predict collec- tion. Finally, the synthesis predicting weighted result is obtained. The experimental results show that the new method is superior to the conventional algorithm.
基金
supported by the National Basic Research program of China(973 program)(613610303)
