Study on the Methods of Detecting Cucumber Downy Mildew Using Hyperspectral Imaging Technology

Abstract

Hyperspectral imaging technology, which can integrate the advantages of spectral detection and image detection, meets the need of detecting the cucumber diseases fast and nondestructively. In this paper, hyperspectral imaging technology is adopted to detect the cucumber downy mildew fast and nondestructively. Firstly, hyperspectral images of cucumber leaves infected downy mildew are acquired by the hyperspectral image acquisition system. And optimum wavelengths are collected by the principal component analysis to get the featured images. Then the image fusion technology is adopted to combine collected images with the featured images to form new images by pixel-level image fusion. Finally, the methods of the image enhancement, binarization, corrosion and dilatation treatments are carried out, so the cucumber downy mildew is detected. The result shows that the accuracy rate of the algorithm for detecting cucumber disease can reach nearly 90%. Studies have shown that hyperspectral imaging technology can be used to detect cucumber downy mildew.