基于模糊支持向量机和D-S证据理论的钨矿石初选方法

发布时间：2018-05-09 14:06

本文选题：机器视觉 + 图像处理　；参考：《光子学报》2017年07期

【摘要】：单一特征识别的钨矿石初选准确率低,稳定性差,本文提出结合模糊支持向量机和D-S证据理论相的多特征钨矿石识别方法.对矿石图像预处理后,分别提取矿石的颜色、灰度和纹理等3类视觉特征,对这3类视觉特征进行模糊分类得到各自的信任度,再以这3类信任度为独立证据,采用D-S证据理论对3类证据进行融合,并依据分类判决规则得到最终的识别结果.试验结果表明,通过D-S理论对模糊向量机证据的融合,钨矿石初选的正确识别率达到96%以上,其准确率和稳定性较单一特征均有大幅度提高,满足生产过程中初选工艺的要求.
[Abstract]:The tungsten ore primary selection with single feature is low in accuracy and poor in stability. In this paper, a multi feature tungsten ore identification method combining fuzzy support vector machine and D-S evidence theory is proposed. After preprocessing the ore image, 3 kinds of visual features, such as color, gray scale and texture, are extracted respectively, and the 3 types of visual features are classified by fuzzy classification. The trust degree of the 3 types of trust is the independent evidence, and the 3 kinds of evidence are fused with the D-S evidence theory, and the final recognition results are obtained according to the classification rules. The experimental results show that the correct recognition rate of the tungsten ore primary selection is above 96% through the fusion of the fuzzy vector machine evidence by the D-S theory, and its accuracy and stability are compared. The single characteristics have been greatly improved to meet the requirements of the primary process in the production process.

【作者单位】：南昌大学机电工程学院;江西理工大学机电工程学院;
【基金】：国家自然科学基金(No.71361014)资助~~
【分类号】：TD954;TP391.41

【相似文献】