摘要
由于蔗糖晶体的镜面透明,煮糖显微观察镜采集出来的图像变为了伪三维的图像,失真,不利于制糖过程的自动监控;针对这类特殊图像探求一种较好的图像预处理方法,首先采用几种常用的图像预处理方法如:灰度调整,照度均匀,直方图均衡,对数变换及中值滤波的方法对伪三维图像进行预处理,再对处理结果运用分割效果较好的Canny算法分割,并将最终结果进行比较分析,经比较,采用灰度调整和中值滤波的预处理的图像,分割效果较其它方法理想,分割结果适用于煮糖过程的自动监控。
Because the sugar crystal is smooth and transparent, the imagecollectedty microscope turns into false three-dimensional image and fuzzy, it goes against automatic monitoring in making sugar. Aimed at this special image, a better method is discussed. At first, adopt many methods which are often used to pretreat them, such as light equalization, histogram equalization, logarithm transform and median filter, and then segment the result using Canny algorithm whose result is better. At last, compare to the final result. By comparing, the best pretreatment methods are gray adjusting and median filter, and their results are fit for automatic monitoring.
出处
《计算机测量与控制》
CSCD
2006年第8期1027-1029,共3页
Computer Measurement &Control
基金
广西自然科学基金项目(桂科自0135059)
广西科技攻关项目(桂科攻0322012-5)
关键词
伪三维图像
灰度调整
照度均匀
直方图均衡
中值滤波
false three-dimensional image
gray adjusting
light equalization
histogram equalization
median filter