基于颜色的微血管参数测量研究

针对肿瘤病理学研究和检验中缺乏微血管参数测量手段的状况，采用数字图像处理的方法，对微血管特征参数的测量作了研究．采用了基于色度学的真彩色图像分割算法，以HSI彩色模型为基础，辅以I1I2I3和U^*V^*W^*彩色模型中12和U^*分量，组成五维特征分量描述图像；利用多维阈值分割(MDT)法与三维直方图生长法和颜色聚类法结合分割图像；为解决染色缺损问题，采用了可控的二值形态学的膨胀-腐蚀算法补偿边缘缺损；最后采用灰度直方图阈值分割法区分微血管各部位，并利用链码技术实现了对目标图像的滤波去噪和参数测量．并且对病理学中微血管参数测量的定量化、客观化进行了研究，据此研制出了检测软件，填补了国内此方面病理研究手段的空白，具有良好的应用和推广价值。

RESEARCH ON MEASURING MICROVESSEL'S PARAMETERS BASED ON COLOR

In this paper, a method of digital image processing is provided to measure microvessel's parameters in order to solve the lack of means in neoplasm pathology. The paper presents a 5 dimensional color space to describe a color image, which consists of HSI color model, I2 and U * feature value in I1I2I3 and U *V *W * color model. Based on the mulsti dimensional thresholding(MDT), growing in 3D histogram, and color clustering, some algorithms are proposed to get the edge of microvessels. In order to reduce some flaw on microvessel's edge, a dilate erose algorithm to complement the flaw based on mathematical morphology is discussed in detail. Finally, a gray level thresholding approach is used to separate different parts of microvessel, and the microvessel's parameters are achieved through chain code. This paper studies the practice of measuring microvessel's parameters. It provides a modern detection and analysis means for pathology, which has a good ability and is worth generalizing.