多谱遥感物体检测通过选择性跨模态相互作用和聚合
Minghao Cui1, Jing Nie2, Hanqing Sun3
1School of Microelectronics and Communication Engineering, Chongqing University, Chongqing, 400044, China; College of Computer Science, Chongqing University, Chongqing, 400044, China.
概括
这项研究引入了多谱遥感物体检测的新框架. 它通过选择性交互和汇总跨模式信息来增强特征融合,提高准确性和降低计算成本.
科学领域:
- 地质科学和远程传感
- 计算机视觉 计算机视觉
- 人工智能的人工智能
背景情况:
- 多谱遥感对象检测对于环境和灾害监测至关重要.
- RGB和红外数据的有效融合是系统性能的关键.
- 挑战包括捕捉交叉模式的依赖性和在融合过程中抑制噪音.
研究的目的:
- 提出一个新的框架,选择性交叉模式交互和聚合 (SIA),用于改进多光谱遥感物体检测.
- 为了增强有意义的跨模态远程依赖关系的捕获.
- 在特征融合过程中抑制噪音和无关信息,以获得更好的区分质量.
主要方法:
- 拟议的SIA框架由两个组成部分组成:选择性交叉模式交互 (SCI) 和选择性特征聚合 (SFA) 模块.
- 该SCI模块选择性地优先考虑信息的远程依赖性,降低计算成本.
- 该SFA模块使用一个封闭机制来过来自功能融合的噪音和冗余信息.
主要成果:
- 该SIA框架在无人机车辆,M3FD和LLVIP数据集上实现了卓越的检测准确性.
- 在DroneVehicle基准中,拟议的方法比C2Former的表现优于mAP@0.5.5.的2.8%.
- 与现有方法相比,这种方法显示了较低的计算成本.
结论:
- 通过改进特征融合,SIA框架有效地解决了多谱遥感物体检测方面的挑战.
- 选择性相互作用和聚合导致更高的准确性和效率.
- 该方法显示了各种地质科学和遥感应用的巨大潜力.
相关概念视频
01:21Mass Spectrometry: Complex Analysis
2.1K
Mass spectrometry is an important technique for the identification of pure compounds. However, it has some limitations for the analysis of complex mixtures, often due to excessive fragmentation making the spectrum too complicated to decipher. Mass spectrometry can be combined with suitable separation methods in sequence, forming hyphenated methods, which are useful in the analysis of complex mixtures.
GC–MS is a powerful hyphenated method commonly used in forensics and environmental...
GC–MS is a powerful hyphenated method commonly used in forensics and environmental...
2.1K
01:21Tandem Mass Spectrometry
3.0K
Tandem mass spectrometry is a technique that uses multiple mass analyzers in series to obtain a higher selectivity and reduce chemical noise during analyte detection. Instruments with multiple analyzers separated by an interaction cell enable secondary fragmentation and selected study of the fragment ions.Secondary fragmentations occur in the interaction cell and can be induced by various factors. Fragmentation induced by collision with inert gases, such as N2, Ar, He, etc., is called...
3.0K