朱从容
（浙江海洋学院机电工程学院，舟山 316004）

摘要：提出了一种基于机器视觉的鱼苗计数方法。对采集到的鱼苗灰度图像进行分析，通过数据拟合方法建立图像中鱼苗所占像素点数与鱼苗数的关系，并由此对鱼苗进行计数。采用常用的数字CCD摄像头，配合自行编写的图像处理软件，建立鱼苗计数试验系统。试验结果表明，点算图像内45尾、50尾鱼苗时，计数的准确率可达95%以上，计数时间在3 s以内。

关键词：鱼苗计数；机器视觉；灰度图像；图像处理

A novel fries-counting method based on machine vision technique
ZHU Cong-rong
( College of Mechanical Engineering, Zhejiang Ocean University, Zhoushan 316004, China )

Abstract: A novel fries-counting method based on machine vision technique is put forward in this paper. The grey-scale maps of fries are obtained，and the relationship between the pixel numbers of fries in these maps and the numbers of fries, named counting equation, is setup with data fitting method. The number of fries is counted with the fitted counting equation. A commercial CCD camera and self-compiled image processing software are used to setup an experimental fries-counting system, and the counting process is described. The experimental results indicate that the accuracy of counting approaches is above 95% as the numbers of fries in imagine is 45 or 50. The overall computational time of image processing for the present program is in 3 seconds.

Key words: Fries-counting；machine vision； grey-scale map；image processing