Line-Based Simultaneous Detection and Tracking of Triangles

Abstract

In order to capture noncooperative targets at close distance, vision-based target detection and tracking have been used as key technologies with great challenges. This paper proposes a novel triangular component detecting and tracking method. In the detection stage, line segments are first extracted from the regions containing triangular spans. Isolated but similar segments are merged under the distance and angle constraints. Then the endpoint distance and area ratios are used as the measurement to determine possible triangle candidates. The geometric property is introduced to discard false-positive candidates. In the tracking stage, the triangle model is characterized by the visual features and geometric relationships of its line segments. An affine transform matrix is constructed to find the correspondence between models and objects in consequent frames. Finally, a line matching–based tracking strategy is designed. Experiments on synthetic and natural images verify the effectiveness of the proposed method in detecting and tracking triangles.