通过层次特征融合在大型FOV中有效检测多目标的YOLO-GL
Xinyu Chen1, Chengjun Dong2, Tong Cui1
1College of Artificial Intelligence, Shenyang Aerospace University, Shenyang, 110136, China.
这项研究介绍了YOLO-GL,这是工业安全的先进AI模型. 它提高了火焰和烟雾等关键指标的检测,提高了现场监测的准确性和实时性能.
科学领域:
- 计算机视觉
- 人工智能
- 工业安全工程
背景情况:
- 工业建筑工地面临复杂的安全检查挑战,包括大规模的变化和多个对象的检测困难.
- 准确检测关键安全指标,如火焰,烟雾,人员服装和操作行为至关重要,但经常受到当前技术的限制.
研究的目的:
- 提出YOLO-GL,一个增强的物体检测网络,旨在克服工业环境中当前安全检查方法的局限性.
- 在复杂的工业环境中提高检测关键安全指标的准确性和稳定性.
主要方法:
- 开发YOLO-GL,具有并行本地-全球多层融合模块 (C2f_gl),具有多尺度特征表示的注意力机制.
- 实现层次特征融合架构和适应性特征融合混合模块 (MSF&ACS) 进行动态交叉规模的语义优化.
- 使用公共基准和工业场所集合的复合数据集进行广泛的实验.
主要成果:
- 在火焰检测方面,YOLO-GL实现了最先进的性能,mAP@0.5增加了3.5% (70.8%至74.3%),mAP@0.5:0.95增加了2.9% (38.7%至41.6%).
- 该模型的实时处理能力保持在每秒80.59 (FPS).
- 与现有方法相比,在复杂的工业环境中表现出优越的稳定性.
结论:
- YOLO-GL在工业安全监控中的物体检测方面取得了重大进展.
- 拟议的架构为建筑工地实时安全检查提供了有效和强大的解决方案.
- 功能融合和注意力机制的创新有助于在具有挑战性的场景中提高性能.
更多相关视频
08:25Combining Eye-tracking Data with an Analysis of Video Content from Free-viewing a Video of a Walk in an Urban Park Environment
Published on: May 7, 2019
11:19Multimodal Hierarchical Imaging of Serial Sections for Finding Specific Cellular Targets within Large Volumes
Published on: March 20, 2018
相关概念视频
Association Areas of the Cortex
Prefrontal Association Area: This area is located in the frontal lobe and is involved in planning, decision-making, and moderating social behavior. It connects with primary motor areas,...
Multi-input and Multi-variable systems
In the absence...
Collisions in Multiple Dimensions: Problem Solving
A small car of mass 1,200 kg traveling east at 60 km/h collides at an intersection with a truck of mass 3,000 kg traveling due north at 40 km/h. The two vehicles are locked together. What is the...
Depth Perception and Spatial Vision
Super-resolution Fluorescence Microscopy
Collisions in Multiple Dimensions: Introduction
