Multi-touch tracking recognition method and system

Abstract

The invention provides a multi-touch tracking recognition method and system, wherein the method comprises the following steps: respectively acquiring the positions of touch points in adjacent two-frame images; according to the positions of the touch points, calculating the distances between the touch points in the adjacent two-frame images; according to a sequence of touch-point distances ranked from small to large, gradually searching the correlation modes of the touch points in the adjacent two-frame images; respectively calculating the touch-point dispersion of each searched correlation mode until the touch-point dispersion of the next searched correlation mode is greater than the touch-point dispersion of the previous searched correlation mode for the first time, judging the previous correlation mode is the optimal correlation mode. In the method provided by the invention, through carrying out analysis from a global perspective, and calculating the minimum touch-point distance of the mutually-correlated touch points in each correlation mode and the touch-point dispersion in each correlation mode, the associated accuracy and robustness of the touch points in adjacent two-frame images are ensured.

Description

technical field [0001] The invention relates to the technical field of touch control, in particular to a multi-touch tracking identification method and system thereof. Background technique [0002] With the wide application of touch control products, touch control technology has become a hot technology in today's human-computer interaction. The control of many products is gradually replaced by touch control technology, while touch control technology is constantly challenging new application environments and fields. [0003] In the application environment of touch control based on optical images, users generally hope that the machine can support more touch point operations to meet different application requirements, and also hope that the machine can understand user intentions more intelligently, and then provide users with more natural and intelligent services . In the current touch control system, the machine judges the operation that the user wants to perform by analyzin...

AI Technical Summary

Problems solved by technology

This method is simple and has very good real-time performance, and there are many improved versions of the algorithm at present, but this method still cannot avoid the essential defect of the misunderstanding that the closest Euclidean distance is the optimal association; another method is to use the K nearest neighbor algorithm, This method searches for K touch points with the smallest Euclidean distance in the next frame for each touch point in the previous frame, and sorts them according to the distance value, and then associates each touch point with the touch point with the smallest distance in the previous frame, If a touch point in the next frame is in the K-nearest neighbors of multiple touch points in the previous frame, the idea of ​​"minimum distance takes all" is used for priority matching. This algorithm has good real-time performance and better tracking accuracy than the former, but still It is a local optimal method, which does not consider the optimal association as a whole, that is, it still does not judge whether the current association method is optimal from a global perspective.

Images