不对称的对比学习网络通过知识蒸为无服务铁路表面缺陷检测检测缺陷
IEEE transactions on neural networks and learning systems
|October 29, 2024
概括
这项研究引入了一个新的深度学习模型,用于表面缺陷检测 (SDD),有效地使用RGB和深度数据. 提出的方法显著减少模型大小,同时保持高性能,超过其他16种最先进的技术.
科学领域:
- 计算机视觉 计算机视觉
- 深度学习 (Deep Learning) 是一种深度学习.
- 材料科学 材料科学 材料科学
背景情况:
- 表面缺陷检测 (SDD) 在工业应用中至关重要.
- 现有的SDD深度学习模型主要使用RGB数据,忽视了有价值的深度信息.
- 当前的双流模型增加了计算成本和参数数量.
研究的目的:
- 为无轨表面缺陷检测提出一个新的深度学习框架.
- 通过结合深度特征来解决现有的仅RGB模型的局限性.
- 开发一个参数效率高的模型,实现与较大的双流网络可比的高性能.
主要方法:
- 开发了一种双流教师模型 (ACLNet-T),以提取RGB和深度特征.
- 一个单流学生模型 (ACLNet-S) 为参数效率而设计.
- 知识蒸技术,包括对比蒸,多尺度图形映射和注意蒸,用于将知识从ACLNet-T传输到ACLNet-S.
主要成果:
- 拟议的学生模型 (ACLNet-S*) 实现了与教师模型 (ACLNet-T) 相比的性能,参数数量减少了八倍.
- 在NEU RSDDS-AUG数据集上,ACLNet-S*的性能超过了16种最先进的方法.
- 该模型在三个额外的公共数据集中展示了强大的概括能力.
结论:
- 提出的知识蒸方法有效地转移多式联运特征,使一个紧而强大的表面缺陷检测模型成为可能.
- ACLNet-S*为高效准确的工业表面缺陷检测提供了一个有前途的解决方案.
- 该方法强调了将RGB和深度数据与先进的蒸技术相结合的潜力,以提高性能和效率.
更多相关视频
相关概念视频
Equivalent Resistance
In circuit analysis, situations often arise where resistors are neither in series nor parallel configurations. To tackle such scenarios, three-terminal equivalent networks like the wye (Y) (Figure 1 (a)) or tee (T) and delta (Δ) (Figure 1 (b)) or pi (π) networks come into play. These networks offer versatile solutions and are frequently encountered in various applications, including three-phase electrical systems, electrical filters, and matching networks.
Unsymmetric Loading of Thin-Walled Members
Thin-walled members with non-symmetrical cross-sections are vital to engineering structures, offering material efficiency and structural integrity. However, unsymmetrical loading on these members leads to complex stress distributions, resulting in simultaneous bending and twisting can cause deformation or structural failure. The interaction between bending and twisting requires detailed analysis to ensure structural resilience.
The concept of the shear center is crucial in countering the...
The concept of the shear center is crucial in countering the...
Maximum Deflection
When analyzing beams under unsymmetrical loads, such as a train moving on a bridge, it is crucial to accurately determine the points of maximum stress and deflection. The process involves identifying the maximum deflection of the beam, which may not always occur at its midpoint due to the uneven distribution of the load.
The maximum deflection occurs at a specific point, known as point O, where the tangent to the deflection curve is horizontal. To find point O, the slope of the tangent at any...
The maximum deflection occurs at a specific point, known as point O, where the tangent to the deflection curve is horizontal. To find point O, the slope of the tangent at any...
Bus Impedance Matrix
Calculating subtransient fault currents for three-phase faults in an N-bus power system involves using the positive-sequence network. When a three-phase short circuit occurs at a specific bus, the analysis uses the superposition method to evaluate two separate circuits.
In the first circuit, all machine voltage sources are short-circuited, leaving only the prefault voltage source at the fault location. The positive-sequence bus impedance matrix can be determined by solving the nodal equations,...
In the first circuit, all machine voltage sources are short-circuited, leaving only the prefault voltage source at the fault location. The positive-sequence bus impedance matrix can be determined by solving the nodal equations,...
Differential Relays
Differential relays are used to protect generators, buses, and transformers by comparing electrical quantities at different points. When a fault occurs, the difference in current between the two points triggers the relay to operate, opening the circuit breaker. Under normal conditions, the current entering (i1) and leaving (i2) a generator are equal. When a fault occurs, however, these currents become unequal, and the difference current flows in the relay operating coil, causing the relay to...
Application of Nonlinear Inequalities
A nonlinear inequality describes a comparison involving an expression that curves or behaves more complexly than a straight line. These inequalities often appear in forms that include squares, products, or variables in the denominator.To solve such an inequality, one starts by rewriting it so that zero appears on one side. For example, the inequality: can be factored as: This form makes it easier to identify the values that cause the expression to equal zero. In this case, the key values are 3...


