Magnetic Properties-Based Range Image Descriptor

With the development of a new generation of information technology, range images are being used more and more in the fields of image processing, computer vision, and pattern recognition. More concretely, range images play an increasingly important role in autopilot, medical analysis, and quality inspection. As we all know, the range image is composed of a series of three-dimensional (3D) scattered points. Due to limitations of coordinate data, it makes extremely difficult to distinguish the difference between adjacent points. This paper is such an attempt to address this issue. Through the similarity analysis of electron cloud and range image, this paper presents a range image descriptor based on magnetic properties. Benefit from the classical theory of particle motion in magnetic field, Lorentz force, rotation radius, rotation period, and volume ratio parameters are calculated for each point. On that basis, a complete range image descriptor is designed to encode the potential connection of each point from a higher dimension. Experimental results show the discriminating power of the proposed method. Focusing on range image description, the simple structure of our algorithm can effectively improve the efficiency and accuracy of image processing.