相关实验视频

Updated: Jun 30, 2025

Imaging and Analysis for Quantifying Maize (Zea mays) Abiotic Stress Phenotypes

Imaging and Analysis for Quantifying Maize (Zea mays) Abiotic Stress Phenotypes

Published on: March 28, 2025

使用超光谱成像和卷积神经网络与注意模块的非破坏性检测有缺陷的玉米粒

Dong Yang¹, Yuxing Zhou¹, Yu Jie¹

¹Academy of National Food and Strategic Reserves Administration, Beijing 100037, China; National Engineering Research Center of Grain Storage and Logistics, Beijing 100037, China.

Spectrochimica acta. Part A, Molecular and biomolecular spectroscopy

|March 17, 2024

概括

这项研究引入了一种超光谱成像 (HSI) 方法,使用具有光谱和空间注意力的卷积神经网络 (CNN) 来检测有缺陷的玉米核. 先进的CNN模型在识别各种内核缺陷时实现了高准确度的非破坏性.

科学领域:

农业工程农业工程
计算机视觉计算机视觉
频谱学是一种光谱学.

背景情况:

准确检测有缺陷的玉米粒对于在储存期间保持谷物质量至关重要.
传统的缺陷检测方法往往耗时,破坏性或缺乏精度.
超光谱成像 (HSI) 为分析农产品提供了一个有前途的非破坏性方法.

研究的目的:

开发和评估一个高光谱成像 (HSI) 系统与卷积神经网络 (CNN) 结合,用于不破坏性识别有缺陷的玉米粒.
在CNN模型中研究光谱和空间注意力机制的有效性,以改进内核缺陷识别.
将拟议的基于注意力的CNN模型与传统机器学习算法的性能进行比较.

主要方法:

采集了6种类型的玉米核的超光谱成像 (HSI) 数据 (380-1000nm):发芽,受热损坏,受昆虫损坏,成,破碎和健康.
开发和训练了三个CNN模型:CNN-Spl-At (光谱注意),CNN-Spal-At (空间注意) 和CNN-Spl-Spal-At (融合光谱空间注意).
使用光谱,图像和融合特征,将CNN模型与支持矢量机 (SVM) 和极端学习机 (ELM) 进行了比较.

主要成果:

具有注意力机制的CNN模型在内核缺陷识别方面明显优于SVM和ELM模型.
融合的光谱和空间特征为区分内核类型提供了比单个特征更全面的信息.
在CNN-Spl-Spal-At模型实现了最高的准确性,在训练组中达到98.04%,在测试组中达到94.56%.

结论:

提出的CNN-Spl-Spal-At模型,利用HSI和注意力机制,可以有效地检测各种有缺陷的玉米核,而不是破坏性.
这种方法表明了开发先进的非破坏性测试设备的巨大潜力,用于实时评估玉米的质量.
通过注意模块集成光谱和空间信息对于提高HSI应用中的分类准确性至关重要.

关键词:

分类模型的分类模型.卷积神经网络是一种卷积神经网络.有缺陷的玉米粒.超光谱图像超光谱图像超光谱图像空间注意力空间注意力频谱的注意力注意力.

更多相关视频

Cereal Crop Ear Counting in Field Conditions Using Zenithal RGB Images

Cereal Crop Ear Counting in Field Conditions Using Zenithal RGB Images

Published on: February 2, 2019

RGB and Spectral Root Imaging for Plant Phenotyping and Physiological Research: Experimental Setup and Imaging Protocols

RGB and Spectral Root Imaging for Plant Phenotyping and Physiological Research: Experimental Setup and Imaging Protocols

Published on: August 8, 2017

相关概念视频

Light Acquisition

Light Acquisition

In order to produce glucose, plants need to capture sufficient light energy. Many modern plants have evolved leaves specialized for light acquisition. Leaves can be only millimeters in width or tens of meters wide, depending on the environment. Due to competition for sunlight, evolution has driven the evolution of increasingly larger leaves and taller plants, to avoid shading by their neighbors with contaminant elaboration of root architecture and mechanisms to transport water and nutrients.