光谱成像技术在作物病害检测中的应用进展与趋势

Application of Spectral Imaging Technology for Detecting Crop Disease Information: A Review

病害作为影响农作物生长的主要因素之一,平均每年造成农作物产量损失高达12%以上。病害不仅直接导致农作物产量减少,而且也严重降低了农产品的品质,甚至引发食品安全事故。光谱成像作为一种融合图像处理和光谱学的信息获取技术,能同时获取目标的图像信息和光谱信息,从而更直观表达目标的特征。光谱成像技术可以获得图像上每个点的光谱数据,从而实现对作物病害的颜色、形状和纹理特征及光谱特征的分析,具有快速、直观和无损等特点,近些年在作物病害检测领域的应用取得了较大研究进展。综述了近六年来国内外关于光谱成像技术在作物病害检测领域应用的相关文献,分析了光谱成像技术的优势和局限性,重点阐述了光谱成像作物病害检测中关键的第三个技术:(1)光谱图像分割技术,重点分析了四种常见分割算法的优点和适用范围;(2)光谱特征和空间特征提取技术,重点对比了空间特征、光谱特征和二者加权组合对病害信息表达的准确性;(3)检测模型,重点介绍了光谱植被指数和机器学习模型在作物病害检测中的稳定性和前景。最后,根据上述分析展望了光谱成像技术在作物病害检测领域中应用的研究趋势,为相关研究提供全面且系统的参考。

As one of the major factors hindering crops growth,crop diseases make more than 12%loss of crop yield annually.Diseases not only directly reduce crop yields,but also seriously debase the quality of agricultural products,and even cause food safety accidents.Spectral imaging technique is aninformation foraging approach that fuses image processing and spectroscopy.It couldobtain image and spectral information of crop diseases simultaneously and describediseased spots feature intuitively.Spectral imaging technology improves the accuracy and efficiency of crop disease detection because of the advantage of union of imagery and spectrum and has been a hotspot at present research.This paper reviews the related literatures in recent six years,and analyses the advantages and limitations of spectral imaging technology in crop disease detection and focuses on the third key technology of spectral imaging in crop disease detection.The third key technology of spectral imaging in crop disease detection is emphasized:(1)Spectral image segmentation technology,focusing on the advantages and application scope analysis of four common segmentation algorithms;(2)spectral feature and spatial feature extraction technology,focusing on the accuracy comparison of spatial features,spectral features and their weighted combination of disease information expression;(3)detection model,focusing on the stability and prospects of spectral vegetation index and machine learning model in crop disease detection.Finally,this paper prospects the application prospect and research trend of spectral imaging technology in the field of crop disease detection,and provides a comprehensive and systematic reference for related research.