The Identification Of Color Particles Based On The Closest Point Method On The Energy Surface

With the development of technology, digital image processing technology is walking intopeople’s vision, people began using computer image recognition technology to extract theinformation they need from images. However, because of the deficiencies of color imagerecognition method in microscopic particle image, the promotion and application of imagerecognition technology in the field of particle detection and cell detection has been severelyconstrained. Based on this, we propose a method called the identification of color particlesbased on the closest point method on the energy surface. The method is on the base of colormodel theory, aiming at truly achieve the color particles identification. In this method, IUVcolor model is created, identifying the particle outline feature and local features, getting thecontour feature information such as geometry and shape features of particle, and by means ofspecific characteristics of particles to achieve the particle type identification.Color is the response of different frequency spectra in human eyes. Different frequencieshave different colors, their energy is also different. Based on this theory, we designed a newcolor model called IUV color model. We make the different colors changing into differentenergy components, and generate particulate primary energy surface-I plane, and acharacteristic surfaces-δ surface, then segment them through the improved level set methodwhich is based on the CV model, so we can get the contour information through the mainenergy surface, and the local features from the characteristic energy surface, then dividing theresults of the two geometric operations, to get the detailed characteristics of individual particlesof energy surfaces, to recovery the feature information of particles such as color and textureand so on in greatest degree, and finally achieve the particulate species identification purposeswith the closest point method.This method will lead identification methods of color particles from the edge detectionand face recognition to a surface recognition domain, from two-dimensional space-based to athree-dimensional space, ensuring the color information of image will not be lost, improvingthe accuracy of particle identification, solving the problems of color particle identification,providing a new idea and method for the particle identification. The color image recognition software and urine particle analysis software which are developed on the base of theidentification of color particles based on the closest point method on the energy surface willplay an important role in the particle and cell detection, reflecting the important economic andsocial value.