Research On Color Recognition Of Test Paper

Urine test paper is a common test paper that through the test paper’s color change caused by chemical reaction to analysis the component of test specimen. Test paper color recognition using color image acquisition technology and digital image processing technology to realize the collection of images and color classification and significantly increases the speed and accuracy of the test paper color recognition. So It is one of the key technology of urine test automation equipment.On the basis of a large number of literature study, this paper brought out research based on the problems of urine test strip’s color recognize technology in the test image acquisition, segmentation, interfering process and identification algorithms.Carried out a series of work in adjusting and positioning of light source, color correction and recognition, and achieved certain results.The light adjustment. If the light environment of CCD image collection device is different, the collected pictures’ color usually would be changed more or less. We adopt that composed by 21 LED lights which lined up along the circuit board as the light of CCD device. And adjusting light of the 12 LED among of all with a correction algorithm, each paper pad on the strip is guaranteed to receive light evenly.The light adjustment improves the quality of images that collect by CCD device, and reduce the identification error caused by light changes.The position refinement of the test paper pad. The position refinement method of the test paper pad that based on the main hue energy is proposed to eliminate ill effects caused by the test paper pad’s position offset and help to segment the test paper pad accurately.Color correction. The color correction algorithm is used to correct the color of test paper pad pictures who changed due to dark normal urine color. If the dark urine lead to a significant color change of the test paper pad, the color correction for test paper pad pictures is valueless. So we will output result with special explanation,which ensure the reliability of recognition results.Improvement of color recognition algorithm. The micro-grid color recognition method is proposed to make up for the deficiency of the BP neural network method for recognizing color of urine test strip. It improves the recognition rate with less samples needed for modeling.Platform development. During the process of algorithm research, a software platform is designed to assess the feasibility and effect of the improved algorithm.Using microcomputer and CCD color image acquisition device as the hardware platform to collect images and data, the study of algorithm is directly combined with practical application, which ensure the authenticity and validity of the data.