自制着色底板和光谱特征检测竹叶面积

Detection of Bamboo Leaf Area Based on Self-made Colored Base Plate and Spectral Features

贺安琪 ¹李彬 ¹许章华 ²杨远垚 ¹李增禄³

扫码查看

作者信息

1. 福州大学环境与安全工程学院,福州 350108
2. 福州大学环境与安全工程学院,福州 350108;福建省资源环境监测与可持续经营利用重点实验室,福建三明 365004
3. 福建省资源环境监测与可持续经营利用重点实验室,福建三明 365004
折叠

摘要

为快速准确地检测健康与病态竹叶单叶面积,提出一种基于自制着色底板和光谱特征(CBP-SF)的叶面积检测方法.根据叶片光谱特征设计底板,然后利用波段计算、碎片过滤和自适应阈值方法进行图像分割,再根据竹叶大小进行参照物切割,最后统计叶片区域和参考矩形框的像元数并计算叶面积.与随机森林(RF)、最大类间方差法(OTSU)和叶面积仪法(LAM)的对比结果表明:对于健康竹叶的检测效果,CBP-SF＞RF＞OTSU =LAM;对于病态竹叶的检测效果,CBP-SF＞RF＞OTSU＞LAM;对于全样本竹叶的检测效果,CBP-SF＞RF＞OTSU＞LAM.CBP-SF具备检测健康与病态竹叶单叶面积的能力.

Abstract

In order to detect the single leaf area of healthy and diseased bamboo leaves quickly and accurately,a leaf area detection method based on self-made colored base plate and spectral features(CBP-SF)is proposed.The base plate is designed according to the spectral characteristics of the leaf.However,the image segmentation is performed using band calculation,fragment filtering and adaptive thresholding methods,and then the reference is cut according to the size of the bamboo leaf,and finally the number of image elements in the leaf region and reference rectangle are counted and the leaf area is calculated.Comparison with the random forest(RF),the maximum inter-class variance(OTSU)and the leaf area meter(LAM)methods shows that CBP-SF＞RF＞OTSU =LAM for healthy bamboo leaves,CBP-SF＞RF＞OTSU＞LAM for diseased leaves and CBP-SF＞RF＞OTSU＞LAM for the whole sample of bamboo leaves.CBP-SF has the ability to detect single leaf area of healthy and diseased bamboo leaves.

关键词

着色底板/光谱特征/图像分割/竹叶/叶面积检测

Key words

colored base plate/spectral features/image segmentation/bamboo leaf/leaf area detection

引用本文复制引用

基金项目

国家自然科学基金(42071300)

福建省自然科学基金(2020J01504)

福建省资源环境监测与可持续经营利用重点实验室开放基金(ZD202102)

出版年

2024

实验室研究与探索

上海交通大学

实验室研究与探索

CSTPCD北大核心

影响因子：1.69

ISSN：1006-7167

参考文献量16

段落导航