面向作物干旱胁迫诊断的表型成像技术研究进展OA北大核心CSTPCD

Drought stress of crops is an important factor affecting their yield and sustainable agricultural development.Accurate diagnosis of crop drought stress is the basis for improving water resource utilization.Imaging technology can quickly,automatically,non-destructively,accurately acquire and analyze the phenotype characteristics of crops,providing a powerful new tool for crop science research.This paper focuses on the review of phenotype imaging analysis techniques for crop drought stress diagnosis.First,the single imaging technique for crop drought stress diagnosis was introduced,and then we introduced the fusion imaging technique for crop drought stress diagnosis.In the aspect of single imaging technology introduction,firstly,the principles of six phenotype imaging techniques including RGB imaging,3D imaging,near-infrared imaging,hyperspectral imaging,chlorophyll fluorescence imaging and thermal imaging are introduced in this paper,and then we introduced research progress of the single imaging technology in crop drought stress phenotype analysis,beside the research achievements in crop drought stress in recent years were further summarized.Finally,the imaging technology was summarized and prospected at the end of each imaging technology introduction.In the fusion imaging technology of crop drought stress,this paper first summarized the research results of the automatic comprehensive phenotype imaging analysis platform of crop drought stress in recent years,and then summarized and analyzed the research of different fusion imaging methods,analyzed their advantages and disadvantages,and prospected the future research direction of fusion imaging technology in the end.With the horizontal and vertical comparative analysis of the research results of a single imaging technology,we found that RGB imaging technology has the lowest application cost and the most extensive application range,but the lowest accuracy.The 3D imaging solves the problem of crop occlusion in RGB imaging,improves the accuracy,and is widely used in high-throughput phenotype extraction platforms.The information on crop phenotype parameters can be obtained by near-infrared spectroscopy,chlorophyll fluorescence and hyperspectral imaging in a fast and non-destructive way.Among them,the application scope of NIR imaging is limited due to its limited ability to obtain phenotype information and relatively high cost.Chlorophyll fluorescence and hyperspectral imaging are better at obtaining physiological and biochemical parameters of crops,and are widely used in the fusion of imaging methods.Thermal images obtained by infrared thermal imaging are often used to obtain crop physiological parameters,and are also combined with visible light images for phenotype extraction.In the meantime,the method of obtaining crop phenotype by using the fusion of multiple imaging technologies has the advantages of different imaging technologies,which can effectively avoid the defects of a single imaging technology and make up for the deficiencies of obtaining single imaging phenotype parameters,so as to reflect the crop growth status more accurately and efficiently.More accurate feature extraction can be achieved by integrating various image information obtained by various imaging technologies and using artificial intelligence methods for integrated image processing.The use of fusion imaging technology to obtain crop phenotypes will be one of the important directions of crop drought stress phenotypes extraction in the future.Finally,according to the current development situation,future research on phenotypic imaging technology for crop drought stress diagnosis prospects,including the development of new devices and the combination of new artificial intelligence algorithms.