基于高光谱技术的不同病原引起的马铃薯早疫病识别

打开文本图片集

中图分类号：S532文献标志码：A

Abstract： Potato early blight is highly infectious and destructive. To differentiate the types of potato early blight caused by different pathogens， two potato varieties， Xisen 6 and Qingshu 9， were inoculated with the Alternaria solani and Alternaria alternata pathogens. A hyperspectral sorter was used to monitor the disease progression and characteristics on the infected potato leaves. The hyperspectral data was processed using techniques such as multiple scattering correction， standard normal variable transformation， Savitzky-Golay， and first derivative， combined with methods like principal component analysis， competitive adaptive reweighted sampling， and successive projections algorithm for feature extraction. Support vector machine models were established， as well as SVM models optimized using particle swarm optimization algorithm and genetic algorithm for early blight classification and identification. Among the constructed models， the SG-CARS-SVM model showed the best identification performance， with a test sample identification accuracy of 99.72% 5 balancing accuracy and speed of identification. The results indicate that hyperspectral imaging technology has high feasibility for distinguishing potato early blight caused by different pathogens.

Keywords： potato early blight; spectral preprocessing; feature extraction; support vector machine

马铃薯广适性强、易于种植、产量高，是全世界四大主粮作物之一[1-2]。（剩余10316字）