摘要
研究设计了蔗糖煮糖过程中蔗糖结晶颗粒的检测系统,利用图像识别的方法避免传统蔗糖结晶检测中的物理接触,减轻工人劳动强度,使检测结果更加精确,实现计算机控制煮糖结晶过程,有效地促进了蔗糖产业的长期发展。首先利用百万像素CCD和光学显微镜获取蔗糖结晶的颗粒图像,用图像采集卡完成信号的计算机输入,利用形态学等算法预处理待测图像,然后运用基于遗传算法的otsu法分割图像得到二值图像,最后提取特征值计算出颗粒的个数及面积。试验结果表明处理后的图像分割效果很好,能够准确计算出颗粒特征值,实现了蔗糖结晶颗粒的自动检测。
In this paper, a particle detection system of sugar crystallization is designed. Physical contact is avoided by using image recognition, which reduces labor intensity of workers in detection process, and makes the test resuits moreaccurate. It realizes automatically the detection of crystal particle by computer, which promotes the longterm development of sugar industry effectively. At first, the mega pixels CCD and optical microscope are used to catch the crystal particle image of sugar, and then inputt;ng the signal to computer by image acquisition card. Then using morphological algorithm to pretreat the image, and otsu measurement based on genetic algorithm is adopted to segment the image for getting two value image. Finally, the number and size of particles are calculated by using the extracted eigenvalues. With satisfactory results of segmentation, this method is able to calculate accurately the characteristics of particles, and it is a practical detection method.
出处
《计算机仿真》
CSCD
北大核心
2009年第7期294-297,共4页
Computer Simulation
关键词
颗粒检测
图像识别
阈值
遗传算法
形态学
Particle detection
Image recognition
Threshold
Genetic algorithm
Morphology
