PFLO:基于YOLO架构的田间玉米高吞吐量姿势估计模型
Yuchen Pan1,2, Jianye Chang2, Zhemeng Dong2,3
1College of Computer Science and Technology (College of Data Science), Taiyuan University of Technology, Taiyuan, 030024, Shanxi, China.
Plant methods
|April 15, 2025
概括
研究人员开发了PFLO,一种使用YOLO架构的新型玉米姿势估计模型,用于在具有挑战性的田间条件下准确跟踪作物生长. 这种先进的系统通过克服遮蔽和密集种植等问题来改善精确农业,以更好地监测作物.
科学领域:
- 农业科学 农业科学
- 计算机视觉 计算机视觉
- 植物表型化 植物表型化
背景情况:
- 作物姿势是植物生长和健康的关键指标,对农业生产和研究至关重要.
- 由于背景变化,种植密集,遮蔽和形态变化,在野外条件下准确地估计姿势是具有挑战性的.
- 现有的方法与现实世界田间环境的复杂性作斗争,以精确的作物姿势分析.
研究的目的:
- 开发一个端到端的模型,用于在具有挑战性的田间环境中对玉米的构成估计.
- 为了解决当前位估计技术的局限性,特别是关于遮蔽和密集作物的安排.
- 为实时表型分析和自动作物监测创建一个强大的工具.
主要方法:
- 拟议的PFLO (基于YOLO架构的田间玉米定位估计模型),是一个端到端的深度学习模型.
- 开发了一种新的数据处理方法,使用"关键点-线"注释数据库来生成边界框和构成骨架数据,减轻注释偏差.
- 集成的架构增强用于优化特征提取和选择,以处理复杂的现场条件.
主要成果:
- 在五倍验证组 (1,862张图像) 上,PFLO实现了72.2%的姿势估计 (mAP50) 和91.6%的物体检测 (mAP50).
- 与最先进的模型相比,该模型表现出更高的性能,特别是在检测封闭,边缘和小目标方面.
- 成功重建了玉米作物的骨架姿势,在密集的安排和严重的遮中表现出强大的表现.
结论:
- 对于实时表型分析,PFLO在玉米位估计方面取得了重大进展.
- 该模型有效地克服了特定领域的挑战,使作物监测更准确,更可靠.
- 通过提供自动化植物生长评估的强大工具,PFLO为精准农业的发展做出了贡献.
相关概念视频
Light Acquisition
8.4K
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.
8.4K
Calibration Curves: Linear Least Squares
1.2K
A calibration curve is a plot of the instrument's response against a series of known concentrations of a substance. This curve is used to set the instrument response levels, using the substance and its concentrations as standards. Alternatively, or additionally, an equation is fitted to the calibration curve plot and subsequently used to calculate the unknown concentrations of other samples reliably.
For data that follow a straight line, the standard method for fitting is the linear...
For data that follow a straight line, the standard method for fitting is the linear...
1.2K
Force Classification
1.0K
Forces play a crucial role in the study of physics and engineering. They are essential in describing the motion, behavior, and equilibrium of objects in the physical world. Forces can be classified based on their origin, type, and direction of action.
Contact and non-contact forces are two of the most widely used categories of forces. As the name suggests, contact forces require physical contact between two objects to act upon each other. Examples of contact forces include frictional,...
Contact and non-contact forces are two of the most widely used categories of forces. As the name suggests, contact forces require physical contact between two objects to act upon each other. Examples of contact forces include frictional,...
1.0K


