基于彩色图像信息的机采棉加工智能化控制策略研究

项目名称： 基于彩色图像信息的机采棉加工智能化控制策略研究

项目编号： No.51305164

项目类型： 青年科学基金项目

立项/批准年度： 2014

项目学科： 机械、仪表工业

项目作者： 张成梁

作者单位： 济南大学

项目金额： 24万元

中文摘要： 本项目针对棉花加工混级、混轧、浪费严重的现象，通过机理分析、数学建模、数值仿真与实验验证相结合的途径，突破棉花杂质彩色图像分类检测的关键技术，研究基于模糊C均值聚类的棉花杂质彩色图像分割算法，揭示机采棉加工过程中杂质的除杂机理，如棉枝、铃壳、僵瓣、叶屑、杂草、不孕籽及尘杂等，提出根据含杂类型不同实时优化加工参数的精细化机采棉加工方案。针对每台棉机清理设备的性能特点，深入分析皮棉性状参数、棉花综合质量等级、价格结构、生产效率、衣分等因素，建立基于收益最大化的单目标优化模型，作为衡量棉花加工过程整体性能的目标评价函数，解决非线性、多变量耦合、时变和分布参数系统的建模问题。研究基于适应度排序的遗传算法，并对该模型进行全局最优求解，项目的实施为现代棉花加工行业实现机采棉加工工艺参数自适应调整提供科学依据和技术基础。

中文关键词： 机采棉；图像分割；杂质识别；参数优化；智能控制

英文摘要： In accordance with the phenomenons of mixed grade, mixed rolling and serious waste about cotton processing, through mechanism analysis, mathematical modeling, numerical simulation and experimental verification, key technologise of cotton impurity color image classification detection are broken through. Color image segmentation algorithm for cotton impurity based on fuzzy C-means clustering is studied, removing mechanism of impurity from machine-picked cotton such as cotton sticks, bracteole, stiff flap, leaves crumbs, weeds, mote and dust is revealed, and refined machine- picked cotton processing scheme for processing parameters optimization according to types of impurities is proposed. Based on performance characteristics for each machine, considering lint trait parameters, cotton comprehensive quality grade, price structure, production efficiency and gin turnout single objective optimization model based on maximizing revenue, as objective evaluation function of measuring cotton processing overall performance, modeling problem for nonlinearity, multivariable coupling, time varying and distributed parameter system is solved. Genetic algorithm based on fitness sorting is discussed, and global optimal solution for the model is conducted. Project Implementation provide scientific basis and technology fou

英文关键词： machine-harvested cotton；image segmentation；impurity recognition；parameter optimization；intelligent control