| 水稻病害孢子多光谱衍射识别与病害源定位方法研究 |
| |
| 引用本文： | 杨宁,张天纬,张钊源,张晓东,毛罕平,袁寿其.水稻病害孢子多光谱衍射识别与病害源定位方法研究[J].农业机械学报,2023,54(4):250-258. |
| |
| 作者姓名： | 杨宁 张天纬 张钊源 张晓东 毛罕平 袁寿其 |
| |
| 作者单位： | 1. 江苏大学电气信息工程学院;2. 江苏大学农业工程学院;3. 江苏大学流体机械工程技术研究中心 |
| |
| 基金项目： | 国家自然科学基金（面上）项目（32171895）、国家重点研发计划项目（2019YFC1606600-03）、江苏大学农装学部项目（NZXB20200205）、水稻生物学国家重点实验室开放项目（20200302）和湛江市科技计划项目（2021A05235） |
| |
| 摘 要： | 水稻真菌病害主要依赖真菌孢子在空气中进行传播。然而各种水稻病害孢子的形态相近，传统孢子捕捉仪和显微图像法难以对其进行区分。为了能够准确识别目标病害孢子并进行病害源定位，提出了一种水稻病害孢子多光谱衍射识别与病害源定位方法。为了解决传统衍射方法无法识别形态相似的缺点，设计了一种大视场、无透镜的多光谱衍射成像传感器。通过分析病害孢子衍射指纹图谱，解析稻瘟病菌、稻曲病菌孢子多光谱衍射成像特征规律。融合孢子的形态特征和吸收特性，提出指纹分离强度和相对峰差两个特征参数，建立孢子的多光谱衍射识别模型。通过仿真计算实验分析孢子传播规律，耦合环境信息建立孢子传播过程中的扩散模型。在无定向风及有定向风条件下分析孢子的空间分布情况，提出病害爆发源迭代质心定位算法。实验结果表明，本文方法对水稻病害孢子的识别率达到98.5%,对无定向风条件下的定位误差最低为4.9%,对有定向风条件下的定位误差最低为7.1%。
|
| 关 键 词： | 水稻病害孢子 多光谱衍射识别 孢子扩散模型 病害源迭代质心定位 |
| 收稿时间： | 2022/6/26 0:00:00 |
Multispectral Diffraction Identification of Rice Disease Spores and Localization Method of Disease Source |
| |
| YANG Ning,ZHANG Tianwei,ZHANG Zhaoyuan,ZHANG Xiaodong,MAO Hanping,YUAN Shouqi.Multispectral Diffraction Identification of Rice Disease Spores and Localization Method of Disease Source[J].Transactions of the Chinese Society of Agricultural Machinery,2023,54(4):250-258. |
| |
| Authors: | YANG Ning ZHANG Tianwei ZHANG Zhaoyuan ZHANG Xiaodong MAO Hanping YUAN Shouqi |
| |
| Institution: | Jiangsu University |
| |
| Abstract: | Rice fungal diseases mainly rely on fungal spores for airborne transmission. However, the morphology of various rice disease spores is similar, and it is difficult to distinguish them by traditional spore trap and microscopic image methods. To be able to accurately identify target disease spores and locate the disease source, a multispectral diffraction identification and disease source localization method for rice disease spores was proposed. A large field-of-view, lens-free multispectral diffraction imaging sensor was designed to address the shortcomings of traditional diffraction methods that cannot identify morphological similarities. By analyzing the disease spore diffraction fingerprinting, the multi-spectral diffraction imaging characteristic pattern of rice blast and rice curd spores was analyzed. By integrating the morphological characteristics and absorption properties of spores, two characteristic parameters of fingerprint separation intensity and relative peak difference were proposed to establish the multispectral diffraction identification model of spores. The spore propagation law was analyzed by simulation and calculation experiments, and the diffusion model in the process of spore propagation was established by coupling environmental information. The spatial distribution of spores was analyzed under the conditions of non-directional wind and directional wind, and an iterative plasmodial localization algorithm of the disease outbreak source was proposed. The experimental results showed that the recognition rate of rice disease spores reached 98.5%, and the localization error was as low as 4.9% for undirected wind conditions and 7.1% for directed wind conditions. This method can provide a reference in locating the source of crop disease outbreaks. |
| |
| Keywords: | rice disease spores multispectral diffraction identification spore diffusion model iterative plasmacentric localization of disease sources |
|
| 点击此处可从《农业机械学报》浏览原始摘要信息 |
| 点击此处可从《农业机械学报》下载免费的PDF全文 |
|
