中文摘要：

研制基于机器视觉技术的淡水鱼品种在线识别装置。采用CCD彩色摄像头、图像采集卡、光电开关、数据采集卡、输送机、照明箱等部件组成在线识别装置的硬件部分；基于VisualC＋＋6．0平台编写具备淡水鱼图像采集、图像分析、鱼体特征提取、品种识别等功能模块的在线识别软件程序。利用本装置对鲤鱼、鲫鱼、草鱼、鳊鱼等4种大宗淡水鱼进行品种在线识别。结果表明，以BP神经网络作为识别模型，该装置对4种淡水鱼进行识别的平均准确率达到92．50％，检测所需时间平均为1．3S，该装置可以用于淡水鱼品种的在线、快速、准确识别。

英文摘要：

The on-line identification device of freshwater fish species was developed based on machine vision technology. The hardware devices of the on-line identification device were made up of CCD col- orful camera, image acquisition card, photoelectric switch, data ac quisition card, conveyer and lighting box. The software of the on- line identification device had the function module of the image acqui sition of the freshwater fish, image analysis, and feature extraction of the fish body and species identification, which was compiled by Visual C＋＋ 6.0. Four varieties bulk freshwater fish such as carp, crucian carp, grass carp, bream were identified by the device, The results show that the average accuracy rate of the identification device to four varieties freshwater fish used the BP neural network as the recognition model is about 92.50% and the required average detec- tion time is 1.3 s, the device can be used to identify the freshwater fish species on-line, rapid and accurately.