基于土壤特征信息动态实时获取的玉米精准变量播种方法研究

变量播种可根据土壤条件投入最优化的种子量，充分利用地力资源挖掘产量潜力，节约良种，实现种植收益最大化，是目前播种领域最前沿的研究方向。本项目根据玉米高产高效种植需求，提出一种基于土壤特征信息实时获取与播量在线决策的精准变量播种方法：①研究土壤有机质、质地、含水率的光谱响应特性及其在高速和田间振动下的动态变化规律，提出基于光谱校正与补偿的土壤特征信息原位动态获取方法，实现有机质、质地、含水率多维土壤信息的并行、快速获取；②综合历史播量/产量信息、现代农学知识和田间标定试验，构建基于规则归纳推理和深度神经网络的最优播量决策模型，实现基于实时土壤特征信息的播量在线决策；③研究不同播量区域边界判别方法，建立落种点位置广义定位模型和播量调控超前补偿模型，构建基于局域播量综合的精准变量播种方法和控制策略，实现在土壤条件变化时播种量的实时、精确调整。研究成果将为精准变量播种提供理论依据和新的技术途径。

Variable rate seeding can properly and accurately adjust the amount of seeds according to soil conditions to maximize the return, which is the cutting-edge research area of planting technology. This project proposes a precise variable rate seeding method based on real-time acquisition of soil properties and online decision of seeding rate for maize precision planting. Through real-time collection and rapid analysis of soil spectral information, multi-dimensional soil characteristic parameters including soil organic matter, texture, and moisture content will be obtained. Based on multi-dimensional soil characteristic parameters acquired and integration of historical seeding rate and yield data and modern agricultural knowledge, an optimal seeding rate decision model based on rule induction and deep learning will be built, the real-time decision of seeding rate will be realized, and seeding rate information maps which can accurately match the field and dynamically expand will be generated. Based on the seeding rate information maps, the method of discriminating boundaries of different seeding rate zones will be studied, a generalized positioning model of the seed drop point and a compensation model for seeding rate adjustment will be established, an accurate variable rate seeding system based on local soil properties will be built and experimentally verified. And the real-time and precise adjustment of seeding rate when the soil conditions change will be achieved. The results of this research will provide the theoretical basis and technical support for variable rate seeding based on real-time sensors, which is of great significance to improve the intelligence and precision of maize seeding equipment.