A method based on arc template to detect right-angle corner

Right-angle corner detection is widely used in the field of machine vision applications such as image matching, feature point extraction and image analysis. In order to make the right-angle corner detection quickly and accurately, a novel arc template based right-angle detection and screening method is proposed. This method builds the Cartesian coordinate system based on the USAN region and calculates the degrees of arc corresponding to each point in the USAN region respectively. According to the approximate continuity of the arc corresponding to each point in the angle region, an arc template capable of responding to the right-angle corner point is constructed. The nested experiments of the arc template, Susan corner detection and Harris corner detection show that the radiometric template's response to the right-angle corner has a high accuracy, stability and efficiency.