摘要
针对矿物浮选泡沫大小形状不一、光照不均等特点,提出一种基于适应度反馈微粒群优化算法自适应优选谷底检测阈值的泡沫形态特征提取方法。将微粒群算法的惯性权重和加速因子设置为全局最优点适应度的函数,优选谷底检测阈值,采用局部灰度极小选择边界检测模板,根据多角度逻辑规则比较,获取气泡边界,分割结果的统计分析表明泡沫形态特征服从gamm a分布。
Considering the d ifferent size and shape and poor illum ination ofm ineral flotation froth,the mor-phological characteristics extraction of m ineral flotation froth was proposed based on the adaptive valley edge detection of fitness feedback particles swarm optim ization(FFPSO).Having the inertia weight and accelera-tion factors of PSO taken as the globalm inima function to choose best detection threshold,and having the local gray scale m inima based to select edge detection template,and as well as having the multi-angle logical rules compared to detect froth boundary.Statistical analysis of segmentation results shows that the froth morphologi-cal characteristics are gamma d istribution.
出处
《化工自动化及仪表》
CAS
北大核心
2011年第7期785-788,806,共5页
Control and Instruments in Chemical Industry
基金
国家自然科学基金资助项目(60874069
61071176)
国家杰出青年科学基金资助项目(61025015)
关键词
浮选泡沫
形态特征
微粒群
谷底检测
模糊划分熵
flotation froth
morphological characteristics
particles swarm optim ization
valley edge detec-tion
fuzzy partition entropy