基于小波基特征提取的苹果叶部病害检测算法设计  

A Design of Apple Leaf Disease Detection Algorithm Based on Wavelet Feature Extraction

作　　者：李亚文 赵杰 陈月星 LI Ya-wen;ZHAO Jie;CHEN Yue-xing(College of Electronic Information and Electrical Engineering of Shangluo University/Shangluo Artificial Intelligence Research Center 726000,Shaanxi;College of Biopharmaceutical and Food Engineering,Shangluo University,Shangluo 726000,Shaanxi)

机构地区：[1]商洛学院电子信息与电气工程学院/商洛市人工智能研究中心,陕西商洛726000 [2]商洛学院生物医药与食品工程学院,陕西商洛726000

出　　处：《商洛学院学报》2024年第2期26-34,共9页Journal of Shangluo University

基　　金：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)。

摘　　要：针对苹果叶部常见病害实现绿色、无损检测,提出了一种基于SVM和小波基特征提取的苹果叶部病害识别算法。该算法通过对苹果叶片图像进行小波变换,提取出小波系数后,进一步执行小波包变换,再提取出具有代表性的小波基特征,根据每个区域的特征参数,得到一组小波基特征向量,然后通过SVM进行模型训练,使用SVM分类器对不同病害进行分类识别。试验结果表明,基于小波基特征提取的苹果叶部病害识别算法,识别常见五种苹果叶部病害准确率较高,可靠性较好,满足实际生产中对苹果叶部病害无损检测的需求,为绿色、智慧果业提供技术支持。For the green and non-destructive detection of common apple leaf diseases,an apple leaf disease recognition algorithm is proposed based on SVM and wavelet feature extraction.The algorithm performs wavelet transform on the apple leaf image.After extracting the wavelet coefficients,the wavelet packet transform is further performed to extract the representative wavelet basis features.According to the characteristic parameters of each region,a set of wavelet basis feature vectors is obtained.Then,the model is trained by SVM,and the SVM classifier is used to classify and identify different diseases.The experimental results show that the apple leaf disease recognition algorithm based on wavelet feature extraction has high accuracy and good reliability in identifying five common apple leaf diseases,which meets the needs of non-destructive detection of apple leaf diseases in actual production and provides technical support for green and intelligent fruit industry.

关 键 词：小波基特征 SVM训练 特征提取 召回率 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]