Aiming at the problems of low contrast and fuzzy edges of infrared images of leaking gas, this paper proposed a segmentation algorithm for infrared images of gas leakage based on edge enhancement and channel reconstruction. It was optimized on the AP-UNet architecture. Firstly, spatial and channel reconstruction convolution (SCConv) was used as a lightweight strategy to replace the standard convolution operation, which reduced the computational redundancy and improved the detection real-time performance. Secondly, an edge guidance module (EGM) was designed and integrated to accurately capture the edge details of the leaking gas, which ensured that the tiny boundaries of the leaking gas could be accurately recognized and segmented under complex environments and low contrast conditions, thus improved the accuracy and robustness of the detection. In addition, the model used spatial convolution to obtain the inter-frame features of the video, and learned the gas motion law to improve the detection accuracy.The experimental results on the GasVid dataset show that the Dice coefficient, recall rate, and score of the algorithm in this paper reach 77.8%, 94.1%, and 90.8%, respectively, which are 1.4, 3.4, and 0.8 percentage points higher than those of the baseline network. The inference speed reaches 25.2 frames per second, meeting the requirements of real-time detection. The method in this paper can realize high-precision leakage detection, which is better than the detection accuracy of most models, and can be better applied to gas leakage scenarios in complex environments.
SHIRLEYC P, RAJAJ I J, EVANGELIN SONIAS V, et al. Recognition and monitoring of gas leakage using infrared imaging technique with machine learning[J]. Multimedia Tools and Applications, 2024, 83(12): 35413-35426.
[2]
ZHAOQ, NIEX, LUOD, et al. An effective method for gas-leak area detection and gas identification with mid-infrared image[J]. Photonics, 2022, 9(12): 992.
ZUOJinhui, XUWenbin, ZHOUShijie, et al. Gas leakage detection based on spatiotemporal information of low contrast infrared images[J]. Optics and Precision Engineering, 2024, 32(8): 1186-1198. (in Chinese)
[5]
NIEX, ZHANGH, LIUH, et al. An infrared image enhancement algorithm for gas leak detecting based on gaussian filtering and adaptive histogram segmentation[C]//2021 IEEE International Conference on Real-time Computing and Robotics (RCAR), 2021: 359-363.
[6]
ZHAOQ, LUOD, WANGJ, et al. An image enhancement method for gas leak detection based on infrared imaging[C]//2023 IEEE International Conference on Real-time Computing and Robotics (RCAR), 2023: 245-250.
[7]
TIANZ, SHENC, CHENH. Conditional convolutions for instance segmentation [DB/OL]. (2020-07-26)[2024-12-25].
LIQuancheng, CAOJiangtao, JIXiaofei. Infrared imaging gas leak detection method with optical flow enhancement[J]. Journal of Electronic Measurement and Instrumentation, 2023, 37(3): 50-56. (in Chinese)
[10]
XIEL, ZHANGC, YANGC, et al. An infrared gas detection algorithm based on weak supervision[C]//International Conference on Image, Signal Processing, and Pattern Recognition (ISPP 2024), 2024: 1011-1019.
[11]
YANGF, WANGF, LUANZ, et al. Automatic segmentation of gas plumes from multibeam water column images using a U-shape network[J]. Journal of Oceanology and Limnology, 2023, 41(5): 1753-1764.
[12]
LIJ, WENY, HEL. SCConv: Spatial and channel reconstruction convolution for feature redundancy[C]//2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition(CVPR), 2023: 6153-6162.
[13]
ZHANGZ, FUH, DAIH, et al. Et-net: A generic edge-attention guidance network for medical image segmentation[C]//Medical Image Computing and Computer Assisted Intervention-MICCAI, 2019: 442-450.
[14]
BHATTR, UZUNBASM G, HOANGT, et al. Segmentation of low-level temporal plume patterns from IR video[C]//2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops(DVPRW), 2019: 847-854.
[15]
WANGJ, TCHAPMIL P, RAVIKUMARA P, et al. Machine vision for natural gas methane emissions detection using an infrared camera[J].Applied Energy, 2020, 257: 113998.
[16]
OKTAYO, SCHLEMPERJ, LE FOLGOCL, et al. Attention U-Net: Learning where to look for the pancreas[DB/OL]. (2018-05-20)[2024-12-25].
[17]
CHENL C, ZHUY, PAPANDREOUG, et al. Encoder-decoder with atrous separable convolution for semantic image segmentation[C]//Computer Vision-ECCV 2018. Cham: Springer International Publishing, 2018: 833-851.
[18]
GUOM H, LUC Z, HOUQ, et al. Segnext: Rethinking convolutional attention design for semantic segmentation[DB/OL].(2022-09-18)[2024-12-25].