Full
Outline
To address the issues of insufficient generalization ability and low leakage detection accuracy of models under small sample conditions for oil and gas pipelines，a retrospective prototype network（VRPN） based on learnable variational mode decomposition was proposed and applied to the leakage detection of oil and gas pipelines.The end‑to‑end optimization of mode decomposition and dynamic weight allocation were achieved by using learnable convolution kernels and adaptive parameters.Introducing a double‑layer Inception multi‑scale convolutional branch to capture multi‑dimensional signals and extract more precise feature information.Introducing retrospective prototype classification into the historical prototype buffer，the integration of current and historical prototype network to suppress vibration，improve the identification accuracy.A joint optimization objective function of cross‑entropy loss and prototype separation loss was designed to enhance training stability and efficiency.Experiments show that the proposed VRPN network framework performs excellently under limited sample and cross condition scenarios，achieving an average accuracy of 98.50%，which is 0.67% higher than that of ProtoNet.The research can provide a new effective solution for oil and gas pipeline leakage detection.
79
34
| 科 Family | 属数 Number of genus | 种数 Number of species | 占总种数比例 Percentage of total species (%) | 属 Genus | 种数 Number of species | 占总种数比例 Percentage of total species (%) |
|---|---|---|---|---|---|---|
| 鹅膏菌科Amanitaceae | 2 | 11 | 5.26 | 鹅膏菌属 Amanita | 10 | 4.78 |
| 小菇科 Mycenaceae | 2 | 12 | 5.74 | 丝盖伞属 Inocybe | 5 | 2.39 |
| 多孔菌科 Polyporaceae | 8 | 14 | 6.70 | 蜡蘑属 Laccaria | 5 | 2.39 |
| 红菇科 Russulaceae | 3 | 23 | 11.00 | 小皮伞属 Marasmius | 6 | 2.87 |
| 小菇属 Mycena | 11 | 5.26 | ||||
| 光柄菇属 Pluteus | 5 | 2.39 | ||||
| 红菇属 Russula | 17 | 8.13 | ||||
| 栓菌属 Trametes | 5 | 2.39 |
关闭全屏
No. 86 Xueyuan South Road, Haidian District, Beijing
010-62662699(Operation and Technical Support)
qkjq@cast.org.cn
Copyright © 2025 China Association for Science and Technology. All rights reserved. For all open access content, the relevant licensing terms apply.
Sponsored by the Office of the Leading Group for Cybersecurity and Informatization of CAST, and supported by Science and Technology Review Publishing House
