Intelligent 3-D recognition system considering the geometric characteristics of plants and structures in agricultural production systems

考虑农业生产系统中植物和结构几何特征的智能3D识别系统

• 批准号: 06506002
• 负责人: HASHIMOTO Yasushi
• 金额: $ 24.26万
• 依托单位: Ehime University
• 依托单位国家: 日本
• 项目类别: Grant-in-Aid for Scientific Research (A)
• 财政年份: 1994
• 资助国家: 日本
• 起止时间: 1994 至 1996
• 项目状态: 已结题
• 来源: https://kaken.nii.ac.jp/grant/KAKENHI-PROJECT-06506002/
• 关键词: Virtual Reality; Laser measurement; 3-dimensional image; Computer network; Internet; WWW (World Wide Web); Robot; LAN; 人口現実感; 三次元画像

Until now, plant cultivating systems have been treated as a concentrated constant system even though these are defined as a distributed constant system. The aim of this study is to introduce the idea of a distributed constant approach to plant cultivating systems and then realize those intelligent mechanization by using a 3-dimensional image processing system and artificial intelligence.1 A 3-dimensional measuring system combining a lazier range finder and a robot hand was developed, which allowed a highly accurate measurement of any types of plant and fruit shapes.2 Range images for any types of plants and fruits were obtained using a 3-dimensional measuring system, and then a data base for the plant and fruit shapes were built.3 For real-time measurement, a simple algorithm that enables us to rapidly measure fruits (muskmelon, apple, tomato, cucumber, etc.) was developed based on a polygon technique.4 A data communication system between computers, which are connected with LAN or WAN,was developed. This system can treat a great deal of data such as three dimensional image for fruits and transmit it rapidly from one computer to another ones using a compression technique of the image data.5 Optimal structural and space design for plant cultivating system in plant factories were implemented using a virtual reality technique.6 A new recognition algorithm combining neural networks and expert systems, which allows the shape of any fruits to be correctly recognized, was developed. For this recognition, the data of a 3-dimensional range image was used.7 Some techniques newly developed here will be useful for the mechanization and computerization of plant production systems.

迄今为止，植物栽培系统一直被视为集中常数系统，即使这些被定义为分布常数系统。本研究的目的是将分布常数法的思想引入植物栽培系统中，并利用三维图像处理系统和人工智能实现植物栽培系统的智能化。1开发了一种结合懒惰测距仪和机器人手的三维测量系统，其允许对任何类型的植物和果实形状进行高度精确的测量。使用三维测量系统获得任何类型的植物和果实的距离图像，3.针对水果（甜瓜、苹果、番茄、黄瓜等）的实时测量，提出了一种简单的算法，可以快速测量水果的形状。4.开发了一个计算机之间的数据通信系统，该系统通过局域网或广域网连接。该系统能够处理大量的数据，如水果的三维图像，并利用图像数据压缩技术将其快速地从一台计算机传输到另一台计算机。5.利用虚拟现实技术实现了植物工厂中植物栽培系统的结构和空间优化设计。6.提出了一种新的结合神经网络和专家系统的识别算法，可以正确识别任何水果的形状。为了进行这种识别，使用了三维范围图像的数据。7这里新开发的一些技术将有助于植物生产系统的机械化和计算机化。

项目成果

H.Nishina: "Identification of plant production system based on neural networks" Proc.First Asian Control Conference. 3. 13-16 (1994)

H.Nishina：“基于神经网络的植物生产系统识别”Proc.First Asian Control Conference。

羽藤堅治: "三次元データベースに基づく植物工場のシステム設計(第3報)-仮想植物工場のための距離画像の三次元画像処理-" 植物工場学会誌. 8(1). 43-48 (1996)

Kenji Hato：“基于 3D 数据库的植物工厂系统设计（第 3 次报告）- 虚拟植物工厂的距离图像的 3D 图像处理”植物工厂学会杂志 8(1) (1996)。

T.Morimoto: "Simulative estimation for the identification of the cumulative response of a plant using neural networks" Journal of Society of High Technology in Agriculture(植物工場学会誌). 8(2). 42-48 (1996)

T.Morimoto：“使用神经网络识别植物累积响应的模拟估计”农业高科技学会杂志 8(2) (1996)。

Y.Hashimoto, T.Morimoto, and J.De Baerdemaeker: "New approach to total production systems based on an intelligent control" Preprints of 1st IFAC/CIGR/EURAGENG/ISHS Workshop on Control Applications in Post-Harvest and Processing Technology. 151-156 (1995)

Y.Hashimoto、T.Morimoto 和 J.De Baerdemaeker：“基于智能控制的总体生产系统新方法”第一届 IFAC/CIGR/EURAGENG/ISHS 收获后和加工技术控制应用研讨会的预印本。

K.Hatou,: "Computerized Diagnosis System for Plant Growth in Plant Factory." Proc.First Asian Control Conference. Vol.3. 9-12 (1994)

K.Hatou，：“植物工厂植物生长计算机化诊断系统”。