Segment test based X corner detection and sub-pixel localization

In order to achieve the ability of rapidly detecting and accurately locating X corners, a real-time method to locate X corner is proposed. Firstly, the image is filtered to find X corner candidates according to symmetry and intensity distribution characteristics of X corner image, and then the segment test is performed for each X corner candidate to confirm it's a X corner, find edges and localize corner in sub-pixel. Experiments show that compared with Harris algorithm, the method can detect X corners and eliminate other corners, provides better positioning accuracy and higher detection speed, and has the capability to process PAL images online and in real time.