多组件梯度增强器,用于准确地检测和量化叶面上的
Huan Song1, Lijun Wang2, Yongguang Hu3
1School of Mechanical Engineering, North China University of Water Resources and Electric Power, Zhengzhou, 450045, China. songhuan@ncwu.edu.cn.
Scientific reports
|October 6, 2025
概括
MCGE-Frost通过增强图像梯度来准确检测叶子,减少精准农业中的细分错误. 这种方法提供了高效的实时冰监测,以改善作物管理.
科学领域:
- 农业科学 农业科学
- 计算机视觉 计算机视觉
- 图像处理 图像处理
背景情况:
- 准确的叶子探测对于农业监测至关重要.
- 由于复杂的背景和微妙的叶纹理,现有的方法面临细分挑战.
研究的目的:
- 开发一种先进的方法,精确检测叶子.
- 在复杂的环境中克服当前技术的局限性.
主要方法:
- 拟议的MCGE-Frost (多元件梯度增强) 方法.
- 整合了色彩空间分析 (HSV,实验室) 与渐变融合理论.
- 采用自适应加权和形态过来提高对比度和降低噪音.
主要成果:
- MCGE-Frost实现了3.29%的细分错误率,超过了ExG (8.63%),OTSU (8.98%) 和HSV (11.98%).
- 与深度学习方法相比,计算复杂度降低了40%.
- 在GPU加速系统上实现0.8秒/图像处理.
结论:
- MCGE-Frost提供了一种强大而高效的解决方案,用于实时检测叶子结.
- 该方法增强了用于精密农业的量化智能.
- 支持用于防和作物管理的可操作见解.
相关概念视频
Light Acquisition
9.4K
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.
9.4K
Frost Circles for Different Conjugated Systems
3.6K
The inscribed polygon method is consistent with Hückel’s 4n + 2 rule and helps to learn whether the given cyclic compound is aromatic or not. The compound is stable and aromatic if every bonding molecular orbital (MO) is completely filled with a pair of electrons. However, if the non-bonding or antibonding orbitals are filled with electrons, the compound is unstable and not aromatic. Consider the Frost circle diagrams for cycloalkenes containing 4 to 8 carbons.
3.6K


