基于DSP的血型图像判读系统  

Blood image recognition system based on DSP

作　　者：罗刚银[1] 王弼陡[1] 孙海旋[1] 程文播[1] 吴冬[1] 

机构地区：[1]中国科学院苏州生物医学工程技术研究所中科院生物医学检验技术重点实验室,江苏苏州215163

出　　处：《液晶与显示》2015年第2期283-289,共7页Chinese Journal of Liquid Crystals and Displays

基　　金：江苏省科技支撑计划项目(No.BE2013637)

摘　　要：为了减少人眼判读对血型分析结果的人为因素干扰,设计了自动血型图像判读系统。对该系统所采用的硬件组成和软件识别算法进行研究。采用DSP芯片、SDRAM芯片、FLASH芯片、图像编码芯片、图像解码芯片、U盘/SD卡读写芯片等设计了本系统的硬件;采用灰度变换、平滑滤波、模板匹配、阈值处理、颜色分量提取等算法设计了本系统的软件识别算法。实验结果表明:本系统采用的算法可以有效地分割出血型卡中的微柱小管,并实现红细胞分布的有效识别;同时,采用颜色分量提取的方法可以有效地区分血型卡的种类。本系统基本满足血型图像自动判读的要求。In order to reduce the human disturbance for blood analysis results,an automatic blood type image recognition system is established,and the system＇s hardware and software recognition algorithm is researched.The hardware of the system consists of a DSP chip,a SDRAM chip,a FLASH chip,an image coding chip,an image decoding chip,a U disk/SD card control chip,et al.The software recognition algorithm consists of gray-scale transformation,smoothing filtering,template matching,threshold processing,color component extraction,et al.Experimental results indicate that the software recognition algorithm can segment the micro-tubular from the blood card and recognize the red blood cell distribution effectively.In addition,the blood card＇s type can be differentiated through the color difference effectively.This system can meet the basic requirement of blood type image＇s automatic recognition.

关 键 词：血型分析 DSP 模板匹配 阈值处理 颜色分量 

分 类 号：TP394.1[自动化与计算机技术—计算机应用技术] TH691.9[自动化与计算机技术—计算机科学与技术]