Modified algorithm research of target tracking of mean shift infrared imagery

摘要

The moving target tracking of infrared imagery has been usually applied in the field of imaging and guidance military reconnaissance of weapon, and intelligent transportation system, video surveillance, etc. for civil use. The application of mean shift in the tracking of infrared target was investigated and the target model was built by utilizing the weighted kernel histogram of grayscale likelihood ratio, and then the disadvantages of mean shift were analyzed and modified in the target tracking. The modified algorithm has better tracking effect by contrasting experiments than the conventional one.