Variable Rate Applications for Agrochemical Inputs per Head of Lettuce

生菜种植者目前在作物生命周期中进行7至8次“干预”：*土壤测试/基础养分(N，P，K)应用*种植-磷(P)应用*VRN试验*除草和氮素(N)应用3周*每周通过预防性农药，总共6至7次应用*二次肥料应用*收获*收获后使用除草剂对于每一次干预，蜂鸟将收集和分析数据，并随后进行试验，以建立能够帮助种植者更有效、高效和有利可图地针对其干预措施的模型。该项目的最终目标是允许种植者根据植物的需要实现每株数据驱动的可变速率应用。每个区域如下突出显示：基础应用：_SOIL_1。田间历史：田间形状文件、数字高程模型(DEM)、以前的作物和土壤质量报告。这将形成用来确定该领域当前质量的‘证据’。应用模式：以田间历史和作物类型为输入，确定理想应用条件下的田间养分需求。指示生成：考虑物理限制(使用的农业机械、可用的化肥等)将理想的应用转换为可变比率的指示地图。植物应用：氮和农药_1.大小和计数图：来自遥感图像的NDVI将与海拔地图、土壤和天气数据相结合，生成理想的应用。关于作物产量和对第一次申请的反应的数据随后将被用来通知随后的申请。应用模型：使用步骤1的结果--在理想的应用中确定田地的养分需求。指示生成：考虑物理限制(使用的农业机械，可用的化肥)，并将理想的应用转换为可变比率的指示地图。农药：抗病品种目前被给予不必要的杀虫剂应用，因为它们与非抗病品种一起种植在田间。基于品种类型的Shapefile应用程序地图将在需要的地方针对这些应用程序。收获：由于植株太小或太大，大部分产量损失发生在收获时。基地和植物应用将致力于减少这种变异性，但持续的大小和计数图将把收获钻机引导到田间区域，由于其大小，将首先准备好，然后到下一个区域。收获后除草剂：根据收获的作物、残留作物和杂草的不同比例，从而减少化学过量喷洒、潜在的径流和对环境的影响。

Lettuce growers currently make 7 to 8 'interventions' in the course of a crop life-cycle listed below;* Soil testing / Base nutrient (N,P,K) application* Planting - Phosphorous (P) application* VRN Trials* Weeding and nitrogen (N) application at 3 weeks* Weekly pass with prophylactic pesticide, 6 to 7 applications in total* Secondary fertiliser application* Harvest* Post-harvest herbicide applicationFor each intervention, Hummingbird will collect and analyse data and subsequently conduct trials to build models that can help growers target their interventions more effectively, efficiently and profitably. The ultimate aim for this project is to allow growers to achieve data driven variable rate applications per individual plant based on the plant's needs.Each area is highlighted below:Base Application: _Soil_1. Field History: Field shapefiles, Digital Elevation Models (DEM), previous crops, and soil quality reports. This will form the 'Evidence' used to determine the current quality of the field.2. Application Model: taking field history and crop type as inputs, determining the nutrient requirements of the field in an idealised application.3. Instruction Generation: considering physical constraints (farm machinery used, fertiliser available, etc) converting the ideal application into a variable rate instruction map.Plant Applications: nitrogen & pesticides_1. Sizing & counting maps: NDVI from remote sensed imagery will be combined with elevation maps, soil and weather data to generate the idealised application. Data on the crop performance and response to the first application will then be used to inform the subsequent applications.2. Application Model: Using the results from step 1 - determining the nutrient requirements of the field in an idealised application.3. Instruction Generation: considering physical constraints (farm machinery used, fertiliser available) and converting the ideal application into a variable rate instruction map.Pesticides: Disease resistant varieties are currently given unnecessary pesticide applications as they are planted in the field with non-resistant varieties. Shapefile application maps based on variety type will target these applications where needed.Harvest: Most yield loss occurs at harvest due to plants that are either too small/big. Base & Plant Applications will aim to reduce this variability but continuations of the sizing & counting maps will direct the harvest rig to areas of the field that, due to their size, will be ready first then onto the next zone.Post-harvest Herbicide: Vary rates according to; harvested crops, residual crops and weeds thereby reducing chemical overspray, potential runoff and environmental impact.