Abstract
Camshift is an adaptive tracking algorithm based on color histogram with the advantages of translation, scale and rotation invariance, which is widely used in visual tracking. But in the moving scenes, due to the changing background, occlusion and environmental illumination, the tracking effectiveness and efficiency will be affected and even lead to loss of moving target. Combined with the idea of both bottom-up and top-down, this paper suggests an improved one by using pixel filtering, histogram improvement and occlusion judgment, as well as Kalman filter. The comparative analysis results on the traditional Camshift algorithm and combined with Kalman traditional Camshift algorithm demonstrate that the proposed algorithm enhances the accuracy and stability in complex condition in an effective way and improves the real-time performance as well. Keywords—Visual Tracking; Camshift Algorithm; Kalman Filter; Color Histogram; Occlusion Judgment
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call