基于多颜色分量的尿常规自动检测

AUTOMATIC DETECTION OF URINALYSIS BASED ON MULTI-COLOR COMPONENTS

针对手机终端进行尿常规自动化检测问题,提出一种基于多颜色分量的尿常规检测试纸图像处理算法。对HSV颜色空间中的S分量以及RGB颜色空间中的各个颜色分量执行阈值分割,实现试纸定位,通过透视变换算法做角度校正;使用颜色直方图与投影法对待检测色块进行分割;基于HSV颜色空间建立标准颜色库,将分割后的图像与标准颜色库中的图像利用欧氏距离进行颜色识别。实验结果表明,该方法可以有效地实现尿液试纸条图像分割与检测。

To solve the problem of automated urine routine test for mobile phone terminals,we propose an image processing algorithm for urine test strips based on multi-color components.Threshold segmentation of S components in HSV color space and each color component in RGB color space were performed to determine the strips location in image,and the perspective transformation algorithm was employed for angle correction;the color blocks to be detected was segmented by using the color histogram and the projection algorithm;the standard color library based on the HSV color space was established,and color recognition was carried out by using the Euclidean distance between the segmented image and the image in the standard color library.The experimental results show that the method can effectively realize the image segmentation and detection of the urine test strips.