Abstract:The instance segmentation model based on deep learning has addressed the challenges in intelligently detecting water leakage in shield tunneling. Due to the limited generalization ability of the baseline model, occurrences of missed detections, false detections, and repeated detections are encountered during the actual detection of tunnel water leakage. This paper adopts Mask R-CNN as the baseline model and introduces a mask cascade strategy to enhance the quality of positive samples. Additionally, the backbone network in the model is replaced with RegNetX to enlarge the model's receptive field, and MDConv is introduced to enhance the model's feature extraction capability in the edge receptive field region. Building upon these improvements, the proposed model is named Cascade-MRegNetX. The backbone network MRegNetX features a symmetrical block structure, which, when combined with deformable convolutions, greatly assists in extracting edge features from corresponding regions. During the dataset preprocessing stage, we augment the dataset through image rotation and classification, thereby improving both the quality and quantity of samples. Finally, by leveraging pre-trained models through transfer learning, we enhance the robustness of the target model. This model can effectively extract features from water leakage areas of different scales or deformations. Through instance segmentation experiments conducted on a dataset comprising 766 images of tunnel water leakage, the experimental results demonstrate that the improved model achieves higher precision in tunnel water leakage mask detection. Through these enhancements, the detection effectiveness, feature extraction capability, and generalization ability of the baseline model are improved. The improved Cascade-MRegNetX model achieves respective improvements of 7.7%, 2.8%, and 10.4% in terms of AP, AP0.5, and AP0.75 compared to the existing Cascade Mask R-CNN model.

The Sensitive Data Leakage Detection Model Based on Bayesian Convolution Neural Network

Abnormal Network Traffic Detection Algorithm Based on Improved Convolutional Neural Network

Network Intrusion Detection Model Based on CNN and GRU

Convolutional neural networks-based valve internal leakage recognition model

Network Intrusion Detection Method Based on CNN-BiLSTM-Attention Model

An intrusion detection system based on convolution neural network

Leakage Detection in Water Distribution Systems Based on Time–Frequency Convolutional Neural Network

Time-Frequency Distribution Map-Based Convolutional Neural Network (CNN) Model for Underwater Pipeline Leakage Detection Using Acoustic Signals

Pratical Privacy-preserving Convolutional Neural Network Inference Framework with Edge Computing for Health Monitoring

Water Pipeline Leak Detection Based on a Pseudo-Siamese Convolutional Neural Network: Integrating Handcrafted Features and Deep Representations

An Enhanced Convolutional Neural Network in Side-Channel Attacks and Its Visualization

Probabilistic Model of Object Detection Based on Convolutional Neural Network

A Method of Containment Leakage Rate Estimation Based on Convolution Neural Network

Data Leakage Attack Via Backdoor Misclassification Triggers of Deep Learning Models

Abnormal Traffic Detection Based on Attention and Big Step Convolution

Masked image: Visually protected image dataset privacy-preserving scheme for convolutional neural networks

A Multi-Channel Contrastive Learning Network Based Intrusion Detection Method

A Bayesian convolutional neural network-based generalized linear model

Data Privacy Protection Model Based on Graph Convolutional Neural Network

Intrusion Detection Using Convolutional Neural Networks For Representation Learning

Intelligent Detection of Tunnel Leakage Based on Improved Mask R-CNN