Research On Classsification And Counting Of The Blood Cell

Automatic blood cell counter is an indispensable tool for clinical examination, and its differential counting ability for red blood cells(RBC) and white blood cells(WBC) is a key technical parameter. When our body suffers some diseases, the number and the shape of blood cells will change, the automatic blood cell counter can detect and quantify such changes. The differential counting research for RBC and WBC has great significance, for clinicians can judge the state and the severity of diseases based on these changes. Given the complexity of the blood analyzer, the harsh conditions of its application envrionment and the high price, its application fields is limited, so the reliable low-cost blood analyzer is the current research focus with great commercial value.This paper focuses on the image processingbased automatic blood cell counter. It introduces system flows and algorithms from the following four respects, the acquistion method, cell image segmentation algorithms, feature analysis of blood cell image and the choosen of cell classification machine, it also compares and analyzes their performance. Finally, it accomplishes a contact imaging based counting system for differente blood cells. The system reduces the cost greatly without lower the accuracy of clinical diagnosis. The main works and research results of this paper are listed as follows:(1) It studies the principle of some blood cell image related segmentation algorithms, including threshold segmentation, region segmentation, edge detection operator, and several specific theories based segmentations, and compares the segmentation quality of collected red blood cells and white blood cells images through simulation.(2) It analyzes the features of blood cells, compares the statistical characteristics, shape, texture features of the red blood cells and white blood cells images, and the characteristics of the similarity measure differences between these images. It also introduces the performance, advantages and disadvantages of some representative classification algorithms, providing a strong theoretical practical basis for the next new classification system.(3) Finally, this paper proposes a new contact imaging based blood cell count system. It designs detailedly from the acquisition of blood cell image, image preprocessing to the final accomplishement of the classification algorithm counting system. Finally, it designs a control panel by C++ language programing to realize the observation of intermediate results and final classification results, the panel contains several buttons, such as the simulator pre-processing, threshold segmatation and cell counting.The research acheivements of this paper, in the application view, have great reference value for the design of simple and portable blood cell analyzer at low cost. Meamwhile, the research methods and application framework of this paper can be applied to more application fields, such as the realization of the counting for other cell thpes and identification of processed specific element.