基于GA-SVM和特征提取的苹果叶部病害识别检测
Detection of Apple Leaf Disease Identification Based on GA-SVM and Feature Extraction
李亚文 1何甜2
作者信息
- 1. 商洛学院电子信息与电气工程学院,商洛 726000;商洛市人工智能研究中心,商洛 726000
- 2. 商洛学院电子信息与电气工程学院,商洛 726000
- 折叠
摘要
准确识别与防治苹果叶部病害,能够有效提高苹果的产量与品质.为了提高苹果叶部病害识别的准确率,以常见的5种苹果叶部病害(斑点落叶病、褐斑病、花叶病、灰斑病、锈病)为研究对象,阈值分割苹果叶部图像,通过Hu矩提取图像的形状特征,利用遗传算法(GA)与支持向量机算法(SVM)相结合模型,对特征参数进行分类训练,以达到准确识别苹果叶部病害的目的.实验过程中,利用样本库中的1000张图像进行实验,其中700份样本用于模型训练,五种叶部病害共选用300份进行模型测试.测试结果表明,在遗传算法优化下的支持向量机预测数据精度明显提高,苹果叶部斑点落叶病、褐斑病、花叶病、灰斑病、锈病识别率分别达到94.5%、86.5%、95.5%、90.0%、93.2%.
Abstract
Accurate identification and prevention of apple leaf diseases can effectively improve the yield and quality of apples.To improve the accuracy of apple leaf disease identification,five common apple leaf diseases(spotted leaf disease,brown spot disease,mosaic disease,gray spot disease and rust disease)were regarded as research objects,the apple leaf image was segmented by threshold segmentation,and the shape features of the image were extracted by Hu moments.The genetic algorithm(GA)and support vector machine(SVM)were combined to classify and train the feature parameters,to identify the apple leaf diseases accurately.During the experiment,1 000 images from the sample library were used for the experiment,of which 700 samples were used for model training,and 300 samples were selected for model testing for five leaf diseases.The test results showed that the accuracy of the support vector machine prediction data optimized by the genetic algorithm was significantly improved.The recognition rates of apple leaf spot leaf,brown spot,mosaic disease,gray spot and rust disease were 94.5%,86.5%,95.5%,90.0%and 93.2%,respectively.
关键词
遗传算法/支持向量机/形状特征/苹果叶部/病害识别Key words
genetic algorithm/support vector machine/shape features/apple leaf/disease identification引用本文复制引用
基金项目
陕西省科技厅科技计划项目(2023-JC-QN-0661)
商洛学院科研创新团队(19SXC03)
2021年陕西省本科高等教育教学改革项目(21BY162)
出版年
2024
NETL
NSTL
万方数据