Improved Object-Tracking Algorithm for an Underground Mining Environment

The continuously adaptive mean shift (CAMshift) algorithm, an adaptive tracking model based on colour probability distribution, is effective for coping with simple cases. However, it tends to fail in the environment of a coal mine, in which complex backgrounds and low contrasts occur. This paper proposes an improved CAMshift model which generating a linear combination of the hue and saturation component to obtain state-of-the-art tracking results regardless of non-uniform illumination, similar color interference and low-level distinctiveness. In addition, speeded up robust features (SURF) is incorporated to relocate the search window when tracking is lost. The proposed object-tracking algorithm using SURF and improved CAMshift model is named as SI-CAMshift. Rigorous experiments are conducted to demonstrate the effectiveness of the object-tracking algorithm both in stability and repeatability. Moreover, real-time tracking is a further advantage of the model.