Improved classical Hough transform applied to the manhole cover's detection and location

The lost alarm of sewer manhole cover is very important in municipal administration. An improved classical HT (Hough transform) is proposed for the complex background of image. On binary edge image, all the contours are found using contour tracking. Afterwards, contour filter is designed to eliminate inveracious targets and HT is used on every contour neighborhood in contour chain. This scheme has the advantages of high speed and solves the problem of useless sampling and accumulation. The experimental results show that this approach can get acceptable results in image with complex background, different angles of view, various conditions of sunlight and partly covered phenomenon.