A cavity detection method based on machine vision

In order to search and rescue the victims in rubble effectively, a 3D model of the cavity is required. This paper presents a cavity detection method based on machine vision for disaster rescue robot. Laser ring projection technique is developed to lighten the profile of cross-section of the cavity and draw a line around the internal surface. A CMOS camera is used to capture the images which consist of essential profile information as well as various noises. The method of image acquiring and image preprocessing is also described. However, due to the inevitable influence of the harsh environment and lighting conditions, the profile in the acquired image might be partitioned into several segments. Thus, a method of detecting the curve segments based on vector argument is proposed. The 3D model of the cavity is reconstructed by using a series of complete and accurate cross-section profiles which are fitted by the improved cubic B-spline interpolation. Experimental results show that the proposed profile tracking algorithm is effective and adaptable.