Spatio-temporal Context Tracking Algorithm Based on Retinex-enhanced Gray Information and Color Information
Zhang Hongying and Hu Wenbo
(College of Electronic Information and Automation, Civil Aviation University of China, Tianjin 300300)
Illumination variation may result in extreme changes of local or overall image intensity, thus leads to tracking failure. To solve this problem, a spatial-temporal context (STC) tracking algorithm combing with color information and Retinex-enhanced gray information is proposed. Firstly, by comparing and analyzing the sin-gle-scale Retinex algorithm and multi-scale Retinex algorithm, the multi-scale Retinex algorithm is used to enhance image gray level to reduce illumination change influence on image gray level. Then, on this basis, by comparing the color features of various visual models, target model is adopted by hue information. Target tracking model is combined with model enhancement and multi-scale Retinex gray enhancement model. Finally, experimental results show that the proposed algorithm has a higher success than original algorithm. For example, algorithm success rate is about 95% in the Shaking scenario, which is about 24% higher success rate than STC. In comparison with other algorithms, it has better performance and higher tracking accuracy.