Method of Surveillance Video Retrieval Based on Image Content in HSV
Shi Yunjin · Video Engineering · 2015
With the advancement of building Peace City activities, a large number of HD(High Definition) and digitalmonitors has been gradually spread all over the streets, followed by lots of surveillance videos rapidly. How tostorage, manage and use the vast amounts of surveillance videos effectively becomes a research hotspot. Videoretrieval technology based on the content and semantic becomes a feasible method to search information from hugevideos, and attracts more and more scholars and researchers' attention. In this paper, color information of images inHSV space with non-uniform quantization method is utilized and local binary pattern ia used as content based videoretrieval image features. Besides, a novel but easy method is proposed to measure the similarity of two images' histogram to retrieval the target image and the result conformed to the people's subjective visual perception better.