Based On The Fuzzy Similar Matrix Glycosuria Test Paper Color Recognition Research

With the improvement of people’s living standards, there are more and more people who are suffering from diabetes.So, to find a method that can timely, economically and conveniently monitor diabetes urine sugar level has become an urgently solved problem. In medicine, there is a kind of method by visualizing the change in urine sugar test strips’ color with soaked in urine to monitor urine sugar. However, sometimes human eyes observation’ error is relatively large. With the development of computer technology and the upgrading of image capture device, we can use the camera to shoot the color-changed test strips and standard colors in the test tube, and through the computer program automatically identifying the color change in the test strips, to judge whose standard color is close to the test strips, to reduce the error of the human eyes, and thus achieve the purpose of monitoring urine sugar level. In this regard, in view of the above problems, the paper mainly carries out the following learning and research work:(1) With regard to remarkable span characteristics for the captured color of test strips and tubes, during image segmentation, using the Lab color space, the image is segmented individually and merged together in a, b space, thereby to obtain a binary image containing a target region.(2) Firstly, using the mathematical morphology method to remove the smaller noise in the binary image, and then using the connected domain area of law as well as the connected domain minimum bounding rectangle side length method to remove binary image large noise to get the final target binary image.(3) Combining the center coordinates of the connected domain, using the knowledge of mathematics slope and variance, and judging the points on the straight line which doesn’t include the standard color determined, moreover, classifying the test strips are prone to identify and the standard colors.(4) Using HSV color space model, combining the13methods of setting up fuzzy similar matrix in fuzzy mathematics, then to identify special part and fuzzy similarity coefficient of6standard colors. Through the contrast coefficient, the standard color which the biggest coefficient corresponds with is quiet close color of identification test strips.(5) Combining the image processing with experimental data analysis, summarizing the principles of choosing color identification method, and drawing the maximum and minimum values method or arithmetic average method as the main method of identification, Gram-distance method as the supplementary method of identification, which is the conclusion, only by the way the color recognition could achieve the desired effect.