Adaptive scale target tracking method based on structured support vector machine

The invention discloses an adaptive scale target tracking method based on a structured support vector machine, mainly to solve the problems of a large amount of calculation in the case of adaptive scale and tracking in existing video target tracking based on the structured support vector machine. The method comprises the following realization steps: firstly, a structured output support vector machine model is built, and scale variables are added to the model output; then, an image frame with the target already determined is used for updating a decision function; and finally, target tracking is decomposed into coarse tracking and fine tracking, a target position is estimated from a small number of candidate samples and thus a target searching range is narrowed in the case of coarse tracking, and in the case of fine tracking, the target position and the scale are determined on the basis of coarse tracking. Adaptive scale target tracking is realized, the calculation amount in the tracking process is reduced, the real-time performance effects are improved, and the method can be used for determining the accurate position and the real-time scale of the target in video monitoring.