Fast Scene Change Detection in Surveillance Systems Using ANOVA on Color Histogram

Abstract Automated surveillance systems have drawn much attention recently. For a surveillance system to operate and obtain credible sensing of a scenario, visual analysis is the most critical aspect. With an increasing expectation of acceptable outcomes from surveillance systems, light weight but fast visual analyses for producing authentic results are being greatly emphasized. In this paper, we present a new approach for detecting fast scene change in the sequentially captured photographs of a focused location. The system considers the images in RGB color format and analyzes all 3 color layers. In this method, the color value histograms of the images were compared to each other using a statistical procedure called ANOVA. We devised an approach that approximately identifies the locality of change in the scene. Various experimental outcomes based on this visual computation process have also been provided in this paper.