Management Optimization System for Precision Farming Japan model

该项目的目的是开发一个基于土壤肥力和田间作物生长的时空变异性的农业工作管理决策支持系统，考虑如何在日本引入精准农业。研究结果如下：1.1999~2000年，在玉米和白三叶两种不同前茬作物田的420块不同密度的白三叶样地上，对白三叶与杂草的生长竞争进行了调查。结果表明，种植白三叶对所有小区杂草的发生和生长均有抑制作用，存活的白三叶为白三叶，玉米地为白三叶。提出的Lotka Volterra模型较好地模拟了白三叶与杂草生长竞争的动态行为，可分为四种行为模式。利用研制的实时土壤分光光度计采集了0.5公顷水田和0.3公顷旱地的地下土壤反射率。分别从水田和旱地获得了约900和500个数据集，每隔1m的数据扫描间隔，生成了精确的土壤水分、土壤有机质、硝态氮、pH和EC.3参数图。对土壤参数图的时间变化进行了分析，得出了农业生产决策支持系统的管理图，清晰地描绘了土壤参数的稳定高、稳定低和不稳定部分三个管理单元。利用克立格法，在一块0.4公顷的田地中，使用432个处理，在72个地点绘制了小麦和青贮玉米植株的作物生长图。获得了包括株高和SPAD值在内的约10个生长参数。航空多光谱农作物生长图像得到的NDVI图与SPAD图具有较高的相关性。这份技术报告已经发布。

The project has aimed at developing a decision support system for farm work management based on the spatiotemporal variability of soil fertility and growing crops in the field, considering how to introduce precision farming in Japan. Results were as follows.1. Investigation was conducted during 1999 to 2000 on the competition of growth between white clover and weeds at 420 plots with different densities of white clover in two fields of different former crops, com and white clover. Results showed that cultivating the white clover inhibited the emergence and growth of weeds over all plots and that survivals were white clover for white clover plots and white clover for corn plots. A proposed Lotka Volterra model simulated well the dynamic behavior of growth competition between white clover and weeds ; and four behavior patterns could be classified.2. The real-time soil spectrophotometer developed was used for collecting the soil reflectance under the ground in a 0.5 ha paddy and a 0.3 ha upland fields. With about 900 and 500 data sets obtained form the paddy field and the upland field, respectively, at 1 m data scanning intervals, precise soil parameter maps were generated for moisture, soil organic matter, nitrate nitrogen, pH and EC.3. Analysis on the temporal variations in the soil parameter maps derived management maps for a decision support system on farm works, where three management units : stable high, stable low and unstable parts of soil parameters were clearly depicted.4. Crop growth maps of wheat and silage corn plants were produced by kriging technique using 432 treatments at 72 locations in a 0,4 ha field. About ten growth parameters were obtained including crop height and SPAD value. Airborne multi-spectral images of growing crops gave NDVI maps with high correlation with SPAD maps.5. The technical report was published.