Non-Invasive measurement of signals transmitted by Plants using Kalman neuro Texture Analysis

使用卡尔曼神经纹理分析对植物传输的信号进行非侵入式测量

• 批准号: 06556042
• 负责人: MURASE Haruhiko
• 金额: $ 2.69万
• 依托单位: Osaka Prefecture 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-06556042/
• 关键词: Pattern recognition; Measurement by Images; Retina; Finite element method; Poisson's equation; Data contraction; Kalman neuro-computing; Finite element neural network

In a protected plant production system such as a plant factory, the control applications have been limited to environmental controls. The feedback control technology for greenhouse environmental factors such as temperature, humidity, radiation intensity, carbon dioxide concentration and so forth has been developed and successfully implemented. Using the technology, plant growth can be optimized or controlled by adjusting environmental factors. Plants normally respond to change of environmental parameters. For example, stomata activity is sensitive to ambient humidity and CO_2 concentration. Plant tissue rigidity is affected by the availability of water in the root zone.Environmental factors should be controlled based upon the response of plants to them. The development of bio-response feedback control system has been a challenging task for plant production engineers and scientists. Another important aspect is the development of non-invasive technology for acquiring information of growi … More ng plants. The practice of using non-invasive measurements in the plants is essential for the bio-response feedback and/or feed-forward control system of a centrifuge phytotron. No non-invasive method for direct measurement of plant water status, for instance, is currently available. One alternative uses an indirect measurement technique. Those who are skilled in growing plants can sense whether their plants are under adequate water conditions or not, from minor changes in the appearance of their plants through their color and tone, before the plants wilt. It may be possible to predict the leaf water potential from the appearance of plants.Changes in appearance of a plant canopy or a community of plants due to the growth reflect tonal variations over the community of plants. The tonal variation can be transformed into pictorial information electronically in retrieval form. Some image features can be related to the tonal characteristics of the plant canopy that also substantially reflect the plant growth status. Machine vision can be used to monitor plants growth. Continuous capturing images of plants during their life cycle allows monitoring and possibly early detection of defects. To evaluate the growth including the health of plants based on image features of plants obtained by machine vision system, an intelligent information processing system is required that will be able to identify the plant growth stage and diagnose symptoms of stress.In order to achieve the aim of developing such a bio-response feedback control system, the primary concern should be to develop an effective and practical technique for extracting image features which can eventually indicate the plant growth status. In this study, the textural analysis and the finite element retina were tested. The test results showed that problems in implementing the textural analysis are that there is too much flexibility to construct the co-occurrence matrix and the construction of the co-occurrence matrix requires extremely long calculation time.The result also showed that extracted finite element features clearly indicate the change in texture of the captured image due to plant growth. The performance of the finite element retina was better than the textural analysis scheme in terms of resolution and processing time. Less

在诸如植物工厂的受保护植物生产系统中，控制应用已限于环境控制。开发并成功实施了温室环境因子如温度、湿度、辐射强度、二氧化碳浓度等的反馈控制技术。利用该技术，可以通过调节环境因素来优化或控制植物生长。植物通常会对环境参数的变化做出反应。气孔活动对环境湿度和CO_2浓度敏感。植物组织硬度受根区水分的影响，应根据植物对环境因子的反应来控制环境因子。生物反应反馈控制系统的开发一直是植物生产工程师和科学家的一项具有挑战性的任务。另一个重要的方面是发展非侵入性技术来获取生长信息， ...更多信息 ng植物。在植物中使用非侵入性测量的实践对于离心人工气候室的生物响应反馈和/或前馈控制系统是必不可少的。例如，目前没有用于直接测量植物水分状况的非侵入性方法。一种替代方案使用间接测量技术。那些熟练种植植物的人可以在植物枯萎之前从植物外观的微小变化通过它们的颜色和色调来感知他们的植物是否处于足够的水分条件下。从植物的外观可以预测叶水势。植物冠层或植物群落的外观变化反映了植物群落的色调变化。音调的变化可以转换成检索形式的电子图像信息。一些图像特征可以与植物冠层的色调特征相关，其也基本上反映植物生长状态。机器视觉可用于监测植物生长。在其生命周期中连续捕获植物的图像允许监测和可能的早期检测缺陷。为了根据机器视觉系统获得的植物图像特征对植物的生长状况包括健康状况进行评估，需要一个能够识别植物生长阶段和诊断胁迫症状的智能信息处理系统，为了达到开发这种生物反应反馈控制系统的目的，因此，研究一种有效、实用的植物生长特征提取方法是当前植物生长特征提取的主要任务。在这项研究中，纹理分析和有限元视网膜进行了测试。测试结果表明，纹理分析的问题在于共生矩阵的构造过于灵活，而且共生矩阵的构造需要非常长的计算时间。结果还表明，提取的有限元特征清楚地表明了由于植物生长而引起的捕获图像的纹理变化。有限元视网膜的性能优于纹理分析方案的分辨率和处理时间。少

项目成果

Haruhiko Murase: "Speaking Plant Approach in Phytotechnology" Journal of JSAM. 58 (5). 109-114 (1996)

Haruhiko Murase：《植物技术中的植物方法》JSAM 杂志。

H. Murase: "Digital Image Handling by Finite Element Retina for Plant Growth Monitoring" Proc. Int, Conf. Agri. Machniery Eng. Seol Korea. 3. 765-772 (1996)

H. Murase：“用于植物生长监测的有限元视网膜数字图像处理”Proc。

Haruhiko Murase: "Growth Monitoring of Green Vegetables Cultured in A Centrifuge Phytotron." Plant Production in Closed Ecosystems Kluwer Academic Publishers. 305-319 (1997)

Haruhiko Murase：“离心机 Phytotron 中培养的绿色蔬菜的生长监测”。

H. Murase: "Plant Prosuction Systems in Closed Ecosystems" Kluwer Academic Publishers, 343 (1997)

H. Murase：“封闭生态系统中的植物生产系统”Kluwer 学术出版社，343 (1997)

Haruhiko Murase: "Digital Image Handling by Finite Element Retina for Plant Growth Monitoring." Proc.Int.Conf.on Agri.Mach.Eng.3. 765-772 (1996)

Haruhiko Murase：“用于植物生长监测的有限元视网膜数字图像处理。”