尿沉渣图像分割及识别算法研究

Research on Urine Sediment Image Segmentation and Recognition Algorithm

尿沉渣检查中,需要利用一些分析仪器识别尿液中的有形成分,检查尿沉渣中的红白细胞、上皮细胞、管型和结晶等,及时发现人体的肾脏疾病、泌尿道疾病。现阶段,我国常用的自动尿沉渣检测仪相比传统人工镜检查效率更高,尤其在尿沉渣的有形成分识别上准确率更高。基于此,针对尿沉渣的自动识别展开探讨,并结合尿沉渣图像的特点,研究图像分割、增强、选择、特征提取以及图像识别算法。

In urinary sediment examination, it is necessary to use some analytical instruments to identify the constituents in urine, and to examine the red and white blood cells, epithelial cells, tubular and crystalline cells in urinary sediment, so as to discover human kidney diseases and urinary tract diseases in time. At present, the automatic urine sediment detector commonly used in our country is more efficient than traditional artificial endoscopy, especially in the identification of the constituents of urine sediment. Based on this, the automatic recognition of urine sediment is discussed, and the image segmentation, enhancement, selection, feature extraction and image recognition algorithm are studied according to the characteristics of urine sediment image.