旱作马铃薯作物氮素营养快速诊断方法与机理研究

The potato is one of the most important food crops, it is identified as China’s fourth major food. It’s important to improve the yield and reduce the chemical use. In order to satisfy the crop requirement of the field management time, position and reasonable fertilizer rate, the nitrogen diagnosis technique should be applied. The research will conducted on the method and mechanism of rapid diagnosis of potato crop canopy nitrogen content based on the preliminary study results. In order to obtain the effective data in the field, Multi-sensors will be used including multi-spectral camera, spectral radiometer, ultrasonic ranging sensor and GPS. The nitrogen content of potato leaf will be measured in the lab. The multi-phenotyping parameters will be extracted as color, unit cover rate, height, spectral reflectance, vegetation index and sampling location. Firstly, in order to reduce the data interference from soil background filed weed and potato flower, the multi parameters will be corrected with image accurate segmentation and spectrum denoising method. Secondly, the dynamic response of the sensor parameters will be analyzed with the growth stage and nitrogen nutrition changes. The relationship between multi-parameters, different growth stages and nitrogen content will be discussed. As a result, the quantitative indicators of growth stages and sensitive parameter of nitrogen content will be selected. Furthermore, they are used to establish the classification model of potato growth stages and nitrogen content of potato canopy. The key management period of potato (Tuber formation- growth stage) will be tracked, and the field nitrogen content distribution could be shown in GIS map based on the GPS data. The conclusion will be drown with the accurate crop needs and management space-time information. It will provide a rapid and nondestructive method for making precision fertilizer-decision in the potato field.

马铃薯被确定为我国第四大主粮，对其全生长期进行营养诊断实施精细化管理具有重要意义。作为块茎作物，马铃薯冠层氮素营养快速诊断尚存在数据源干扰大、生长动态解析不深入和氮素诊断模型精度不高的问题，进一步开展“旱作马铃薯作物氮素营养快速诊断方法与机理研究”很有必要。为获取有效检测数据源，以近地多光谱成像、光谱辐射仪、超声波测距传感器和GPS多源传感器同步获取田间马铃薯作物表型参数（颜色、单位面积覆盖度、株高、光谱反射率、植被指数和位置信息），提出冠层多光谱图像精准分割和光学信号干扰消噪方法，消减土壤背景、杂草和开花等信息带来的干扰，解析不同生长期表型参数对氮素营养在地上和地下茎间 “峰式”转移的动态响应， 从而建立起适于马铃薯块茎生长阶段追踪的量化指标；根据多源检测参数与生长期马铃薯氮素之间的相关分析，优化氮素敏感检测参数，建立高精度马铃薯氮素营养诊断模型，最终实现旱作马铃薯作物氮素营养的快速诊断

马铃薯是我国第四大主粮，对其全生长期进行营养诊断实施精细化管理具有重要意义。作为块茎作物，马铃薯冠层氮素营养快速诊断生长动态解析不深入、诊断模型在精准化管理中应用不足、田间应用传感器技术水平不高等问题。开展“旱作马铃薯作物氮素营养快速诊断方法与机理研究”，阐明了马铃薯生长期检测参数对氮素动态的响应机理，建立了生长时期分类模型、叶绿素和水分诊断模型。研发了基于光谱分析和WSN的马铃薯作物诊断应用传感器和数据管理服务平台，并对马铃薯结薯品质进行了检测方法研究。可对马铃薯作物全生长期营养诊断、田间精细化管理提供理论和技术支持。研究主要内容和重要成果包括：.1）基于高光成像的马铃薯作物营养快速诊断机理研究。基于高光谱成像和光谱辐射仪采集叶片、植株和冠层不同尺度数据，对叶绿素和水分含量展开了消噪方法、吸收特征、响应动态分析，建立高精度诊断模型后，实现了营养动态空间分布可视化。进而建立了生长期判别模型，可为马铃薯光谱学诊断机和田间适时管理提供理论支持。.2）基于光谱分析和WSN的马铃薯作物诊断应用传感器研发。分别开发了基于光谱成像的作物叶绿素含量诊断应用系统、便携式马铃薯作物氮素快速诊断传感器系统、基于透射光谱的植物叶片水分快速诊断系统、基于WSN的作物叶绿素监测大田应用传感器系统，上述系统分别在马铃薯、玉米和小麦作物上测试，可为田间应用提供技术支持。.3）马铃薯作物生长快速诊断应用系统开发。用于指导田间精细化管理，开发大田马铃薯作物叶绿素含量诊断的应用系统，集成高光谱传感器，形成了马铃薯叶绿素含量诊断可视化软件。进而，开发了作物长势田间分布分析软件开发、远程管理系统平台，可为田间管理提供技术支持。.4）基于机器视觉技术的马铃薯品质检测方法研究。利用光谱分析和立体视觉技术对马铃薯外观形态、体积、病害发生、品种分类等进行了研究，建立了快速检测模型与分类模型，可为未来马铃薯作物产品品质定级提供支持。