| 基于多分类器DS证据理论融合的水果识别研究 |
| |
| 引用本文： | 蔡浩,郭宏亮.基于多分类器DS证据理论融合的水果识别研究[J].中国农机化学报,2021(2). |
| |
| 作者姓名： | 蔡浩 郭宏亮 |
| |
| 作者单位： | 吉林农业大学信息技术学院 |
| |
| 基金项目： | 吉林省科技厅科技支撑计划重点项目(20190301024NY)。 |
| |
| 摘 要： | 针对不同分类器对不同水果种类识别准确率的不均衡问题,提出一种基于多分类器DS证据理论融合的水果识别方法。该研究选择kaggle上fruits360数据集中的5种水果作为研究对象,首先对预处理后的5种水果图像的颜色、纹理、形状特征进行提取,分别选用BP神经网络、K均值、SVM三种分类器,结合被测图像在每种分类器上的识别结果和各个分类器对不同水果的分类准确率,构建基本概率函数(BPA函数),通过DS证据融合规则对分类器融合后对被测图像进行识别。试验结果表明:该方法对5种水果的识别平均准确率为95.2%,总体标准偏差为0.02993,在提高单分类器识别准确率的同时,解决了分类器对各种水果识别的不均衡问题。对10组测试集识别的平均准确率为93.5%,总体标准偏差为0.055,该方法对水果种类的识别更准确和稳定。
|
| 关 键 词： | 水果识别 多分类器 DS证据理论 BPA |
Research on fruit recognition based on multi-classifier DS evidence theory fusion |
| |
| Cai Hao,Guo Hongliang.Research on fruit recognition based on multi-classifier DS evidence theory fusion[J].Chinese Agricultural Mechanization,2021(2). |
| |
| Authors: | Cai Hao Guo Hongliang |
| |
| Institution: | (College of Information Technology,Jilin Agricultural University,Changchun,130118,China) |
| |
| Abstract: | Aiming at the problem of the imbalance in the recognition accuracy of different fruit types by different classifiers,a fruit recognition method based on the fusion of multi-classifier DS evidence theory is proposed.The study selected five fruits in the fruits360 data set on kaggle as the research object.First,the color,texture,and shape features of the five preprocessed fruit images were extracted,and three classifiers of BP neural network,K-means,and SVM were selected respectively.Combining the recognition results of the tested images on each classifier and the classification accuracy of each classifier for different fruits,construct the basic probability function(BPA)function,and use the DS evidence fusion rules to fuse the classifiers to the tested images Recognition.The test results show that the method has an average accuracy of 95.2%for the recognition of five kinds of fruits,and the overall standard deviation is 0.02993.While improving the recognition accuracy of a single classifier,it also solves the problem of uneven recognition of various fruits by the classifier.The average accuracy of the identification of the 10 test sets is 93.5%,and the overall standard deviation is 0.055.This method is more accurate and stable in the identification of fruit types. |
| |
| Keywords: | fruit recognition multi-classifier DS evidence theory BPA |
| 本文献已被 维普 等数据库收录！ |
|
