小麦群体特征图象智能识别研究

Crop population feature indices, such as leaf area index, biomass, total No.of seedling, plant stem,and spike, the degree of seedling in good order, are important for classified management in crop planting. Presently,these characteristics are usually obtained from intensive manual measurements at the cost of labor and time.Thus,the application of advanced management and planting technology based on variation of place,seeding,and time during wheat growth is still limited. Advances in computer imageing can be used as diagnostic aids in advanced crop management practices. In this project,a system with high recognition accuracy for identifying wheat population image feature indexes was estabblished by using the mathods of image process,machine study, neural network,and manual intelligence etal, synthetically. Compared to traditional geodesic results,recognition precision of wheat group features 86.2% by image technology. It is shown that image recognition is feaible for identifying crop population feature indexes from wheat population images.

利用人工智能和多媒体等现代科技手段，将图象识别技术用于小麦群体特征的信息提取和栽培管理。在实现图像处理技术获得小麦群体特征信息，示例式机器学习形成群体图像识别规则及建立群体图像识别知识库的基础上，用原型设计方法建造“小麦高产群体图像智能识别多媒体专家系统”，将对小麦因苗分类指导、动态管理，提高栽培管理水平具有重要意义。.