With the development of Industry 4.0, although some smart meters have appeared on the market, traditional mechanical meters are still widely used due to their long-standing presence and the difficulty of modifying or replacing them in large quantities. Most meter readings are still manually taken on-site, and some are even taken in high-risk locations such as hazardous chemical storage. However, existing methods often fail to provide real-time detections or result in misreadings due to the complex nature of natural environments. Thus, we propose a lightweight network called DAMP-YOLO. It combines the deformable CSP bottleneck (DCB) module, aggregated triplet attention (ATA) mechanism, meter data augmentation (MDA), and network pruning (NP) with the YOLOv8 model. In the meter reading recognition dataset, the model parameters decreased by 30.64% while mAP50:95 rose from 87.92% to 88.82%, with a short inference time of 129.6 ms for the Jetson TX1 intelligent car. In the VOC dataset, our model demonstrated improved performance, with mAP50:95 increasing from 41.03% to 45.64%. The experimental results show that the proposed model is competitive for general object detection tasks and possesses exceptional feature extraction capabilities. Additionally, we have devised and implemented a pipeline on the Jetson TX1 intelligent vehicle, facilitating real-time meter reading recognition in situations where manual interventions are inconvenient and hazardous, thereby confirming its feasibility for practical applications.

This article presents a neural network model based on edge intelligence for seismic event detection. We implemented the model in hardware using a field-programmable gate array (FPGA) to achieve in-situ detection of seismic events at acquisition nodes or edge nodes. We designed and implemented the model, focusing on its suitability for hardware implementation on FPGA, employing an encoder—decoder structure. The encoder incorporates reparameterization and depthwise separable convolutions. During training, a multibranch structure was employed, which was then converted to an equivalent single-branch structure during inference to reduce model complexity and parameters. The features extracted by the encoder were further learned by the bi-directional long short-term memory (Bi-LSTM) network and then fed into the decoder for classification. We evaluated the model using the stanford earthquake data set (STEAD) and observed a 70% reduction in parameters while achieving comparable detection performance to EQTransformer. Furthermore, the model structure is well-suited for hardware implementation on FPGA. Applying this model to edge devices for seismic event detection can effectively minimize redundant data transmission and enable in-situ quality control.

Forest fire has the characteristics of sudden and destructive, which threatens safety of people’s life and property. Automatic detection and early warning of forest fire in the early stage is very important for protecting forest resources and reducing disaster losses. Unmanned forest fire monitoring is one popular way of forest fire automatic detection. However, the actual forest environment is complex and diverse, and the vision image is affected by various factors easily such as geographical location, seasons, cloudy weather, day and night, etc. In this paper, we propose a novel fire detection method called Fire-PPYOLOE. We design a new backbone and neck structure leveraging large kernel convolution to capture a large arrange area of reception field based on the existing fast and accurate object detection model PP-YOLOE. In addition, our model maintains the high-speed performance of the single-stage detection model and reduces model parameters by using CSPNet significantly. Extensive experiments are conducted to show the effectiveness of Fire-PPYOLOE from the views of detection accuracy and speed. The results show that our Fire-PPYOLOE is able to detect the smoke- and flame-like objects because it can learn features around the object to be detected. It can provide real-time forest fire prevention and early detection.

The autonomous navigation technology of INS has always attracted the attention of many scholars. In satellite-denied environment, making full use of the auxiliary sensors with autonomous characteristics can suppress the dynamic error generated by INS in autonomous navigation effectively. However, auxiliary sensors like an odometer (OD) and laser doppler velocimeter (LDV) will cause instability to the navigation system. To solve the above problems, we propose an improved autonomous inerital-based integrated navigation scheme utilizing vehicle motion recognition. The main research work can be summarized as: (1) Non-holonomic constrained velocity model and tri-axis zero velocity model has been established for inertial-based integrated navigation model. (2) Vehicle motion state recognition method has been proposed utilizing the IMU data. As shown in the experimental results, the proposed method can estimate the IMU biases and increase the positioning accuracy effectively as compared with the traditional methods.

Phase picking is a critical task in seismic data processing, where deep learning methods have been applied to enhance its accuracy. While lightweight deep learning networks have been optimized for edge computing devices, there is a lack of networks developed explicitly for embedded systems. This paper presents a seismic phase picking model, a hybrid network integrating convolutional neural networks and Transformer, designed for embedded systems. Optimizing network parameters and computational resources, the model significantly reduces resource consumption while guaranteeing accuracy. It employs a multi-branch architecture. Specifically, the global branch employs a modified self-attention mechanism, effectively extracting global features through shared contextual information. The local branch retains local information from the input features. Such a multi-branch architecture facilitates effective interaction between global features and local details, thereby more efficiently capturing the relationships among features. The model can be configured into variants with different sizes to match various embedded systems. This research evaluated the model using the Stanford Earthquake Dataset, achieving a precision of 99.9% for the P-phase and 99.3% for the S-phase. On Raspberry Pi, the model reduced inference time by 58.1% compared to the earthquake transformer while maintaining comparable detection performance.

The identification and determination of seismic phase arrival times is a critical task in seismic data processing. In recent years, methods based on Convolutional Neural Network (CNN) and Transformer models have been widely applied in this field. While CNNs offer potential in seismic feature extraction, modeling limitations restrict their efficiency for detailed waveform analysis. Conversely, Transformer models, though powerful, are constrained by their quadratic computational complexity. Recent research has shown that the State Space Model (SSM) represented by Mamba can effectively simulate long-range interactions while maintaining linear computational complexity. Inspired by this, we propose the Seismic Phase Picking Mamba (SPPMamba) model. We have designed a novel Conv-SSM module that combines CNN layers’ local feature extraction capabilities with SSM’s long-range dependency-modeling abilities. This enables the model to effectively identify and utilize the temporal variations of seismic signals, enhancing the model’s analytical capabilities for dynamic seismic feature characteristics. To validate the performance of SPPMamba, we conducted experiments on public seismic datasets. Experimental results show that SPPMamba demonstrates superior performance in seismic phase picking. This study aims to lay a research foundation for developing seismic data processing algorithms based on SSM.