支持向量机在斑潜蝇虫害叶片光谱分析中的应用

测量了斑潜蝇虫害叶片的可见光反射光谱。运用图像处理技术计算出虫害叶片的破损率;采用光谱分析方法选出对叶片破损程度敏感的波长点。运用支持向量机(SVM)和多光谱法,建立基于敏感波长的受害叶片反射光谱分类模型并进行分类试验。试验结果表明,多光谱法分类精度为90%,SVM为93.8%(多项式核函数)和96.9%(RBF核函数),高于多光谱法。

Classification Visible Spectra of Leaf miner-infected Leaves by Support Vector Machine

Leafminer is a vegetable insect pest. The visible spectral reflectance of the leafminer (Liriomyza sativae Blanchard) infected plant leaves was measured. Damaged degrees (DD) of leaves were worked out with the image processing technology, and the sensitive wavelengths related to them were also selected via spectroscopy analysis. Using the support vector machine (SVM) and the multi-spectral method, the spectral classifying experiments were done and the classifying models were set up to recognize the leaf spectra with different DD. The results that used the SVM and the multi-spectral methods were contrasted. The results showed that the classifying precisions of the SVM are 93.8% (using the polynomial-based kernel function) and 96.9% (using the RBF kernel function), which excelled to the multi-spectral method (whose classifying precision is 90%).