The current manual reading of substation pointer meters wastes human resources, and existing algorithms have limitations in accuracy and robustness for detecting various pointer meters. This paper proposes a method for recognizing pointer meters based on Yolov7 and Hough transform to improve their automatic readability. The proposed method consists of three main contributions: (1) Using Yolov7 object detection technology, which is the latest Yolo technology, to enhance instrument recognition accuracy. (2) Providing a formula for calculating the angle of a square pointer meter after Hough transformation. (3) Applying OCR recognition to the instrument dial to obtain the model and scale value. This information helps differentiate between meter models and determine the measuring range. Test results demonstrate that the proposed algorithm achieves high accuracy and robustness in detecting different types and ranges of instruments. The map of the Yolov7 model on the instrument dataset is as high as 99.8%. Additionally, the accuracy of pointer readings obtained using this method exceeds 95%, indicating promising applications for a wide range of scenarios.

Plant species recognition is a critical step in protecting plant diversity. Leaf‐based plant species recognition research is important and challenging due to the large within‐class difference and between‐class similarity of leaves and the rich inconsistent leaves with different sizes, colors, shapes, textures, and venations. Most existing plant leaf recognition methods typically normalize all leaf images to the same size and then recognize them at one scale, which results in unsatisfactory performances. A novel multiscale convolutional neural network with attention (AMSCNN) model is constructed for plant species recognition. In AMSCNN, multiscale convolution is used to learn the low‐frequency and high‐frequency features of the input images, and an attention mechanism is utilized to capture rich contextual relationships for better feature extraction and improving network training. Extensive experiments on the plant leaf dataset demonstrate the remarkable performance of AMSCNN compared with the hand‐crafted feature‐based methods and deep‐neural network‐based methods. The maximum accuracy attained along with AMSCNN is 95.28%.

The unsupervised anomaly detection task based on high-dimensional or multidimensional data occupies a very important position in the field of machine learning and industrial applications; especially in the aspect of network security, the anomaly detection of network data is particularly important. The key to anomaly detection is density estimation. Although the methods of dimension reduction and density estimation have made great progress in recent years, most dimension reduction methods are difficult to retain the key information of original data or multidimensional data. Recent studies have shown that the deep autoencoder (DAE) can solve this problem well. In order to improve the performance of unsupervised anomaly detection, we propose an anomaly detection scheme based on a deep autoencoder (DAE) and clustering methods. The deep autoencoder is trained to learn the compressed representation of the input data and then feed it to clustering approach. This scheme makes full use of the advantages of the deep autoencoder (DAE) to generate low-dimensional representation and reconstruction errors for the input high-dimensional or multidimensional data and uses them to reconstruct the input samples. The proposed scheme could eliminate redundant information contained in the data, improve performance of clustering methods in identifying abnormal samples, and reduce the amount of calculation. To verify the effectiveness of the proposed scheme, massive comparison experiments have been conducted with traditional dimension reduction algorithms and clustering methods. The results of experiments demonstrate that, in most cases, the proposed scheme outperforms the traditional dimension reduction algorithms with different clustering methods.

1 Introduction In recent years, in the field of smart agriculture, many advanced technologies and data-driven approaches have been widely applied to crop yield estimation, soil and water conservation, pest and disease detection and severity evaluation, species classification and farmland ridge segmentation, and have achieved remarkable results. This special issue highlights several issues that still need further research and discussion in smart agriculture, such as Agricultural UAV (AUAV), Agricultural Remote Sensing (ARS), Agricultural Internet of Things (AIoT), Agricultural Artificial Intelligence (AAI), Agriculture Digital Twin (ADT), Deep Learning and their combinations. 2 AUAV and ARSAUAV and ARS can capture field surface information, provide a wide range of agricultural monitoring, and help farmers and agricultural managers make better decisions. In recent years, combining AUAV, ARS and AI have been widely applied to real-time monitoring and analysis of crop growth status, crop health status and growth stage, and assessment of the impact of natural disasters, such as droughts, floods and storms, and help farmers understand the field situation and develop responses and reduce risks in time. Garofalo et al.(2024) highlighted the potential of combining AUAV, ARS and AI in precision agriculture, with the goal of effectively monitoring physiological parameters. By analyzing AUAV and ARS data, farmers can optimize fertilization, irrigation, and pest control strategies to improve crop yield and quality, reduce waste and promote sustainable agricultural development. The application of AUAV, ARS and AI technologies is constantly developing, and combined with big data and deep learning, will provide more powerful support for agricultural production in the future. Jafar et al. (2024) reviewed the methods, applications and limitations of AAI and AIoT in crop disease detection, provided detailed steps for crop disease prediction using AAI and AIoT, and discussed the application of machine learning and deep learning in crop pest detection, and pointed out the future research prospect of combining AAI with intelligent AUAV and AIoT for field disease detection and monitoring. 3 AIoT and Mobile Devices (MD)AIoT and MDs aim to combine AIoT and MD with AI to provide solutions for smart agriculture. Various AIoT sensors are installed in the field to collect soil moisture, temperature, humidity, crop health, weather conditions and other data in real time, and transmit the collected data to the cloud platform to achieve optimal irrigation, fertilization, disease and pest control, and accurate resource management. MDs such as smartphone and AUAV are used to capture crop images in the field to realize real-time monitoring of farmland, including crop variety classification, crop disease severity estimation and pest detection (Pan et al., 2024; Bedi et al., 2024). The widespread application of AIoT and MDs is transforming the agricultural landscape, making it more efficient, sustainable, and able to meet the challenges of smart agriculture. Li et al.(2024a) introduced a variable direction irrigation decision-making method of east-west ridge based on the AIoT management control system. This study provides a new irrigation decision-making way to improve the efficiency of crop production. Li et al. (2024b) summarized the latest development of AAI and AIoT, and proposed an efficient deep learning architecture based on Mobile Vision Transformer (MobileViT) for real-time detection of crop disease. It is designed with high-accuracy and low-cost, making it suitable for deployment on MDs with limited resources. Qiu et al.(2024) proposed a method for measuring water content of millet by using AIoT differential capacitance sensor. It provides a reliable means for the accurate determination of crop water content, and provides a strong support for improving agricultural production efficiency and resource utilization.4 Agricultural Digital Twin (ADT) Digital twins have been applied to various fields, including smart agriculture. ADT is one example of many digital farming innovations. It relies on the integration of state-of-the-art agricultural technologies, including big-data analytics, AAI, AIoT, ARS, AUAV, Information & Communication Technology, Geographic Information Systems (GIS). It helps farmers monitor crop growth, soil conditions and climate change in real time, optimize the use of water, fertilizers and pesticides, reduce resource waste and environmental impact, predict potential risks, such as pests and diseases, climate change, and take steps to reduce losses, create a comprehensive basis for data-driven policy decisions, and promote sustainable development of agriculture. Recently, ADT has promoted the application of new technologies such as AIoT, AAI and big-data analytics to smart agriculture, and become an important tool for modern agricultural management. The application of ADT will revolutionize agricultural production, improving overall efficiency and sustainability. Garske et al. (2024) reviewed "digital twin sustainable development", "digital twin", "digital twin agriculture", "Data governance", "Destination Earth", "EU Data Strategy", and "EU Data Governance Act". By collecting and analyzing large amounts of data, ADT can provide scientific basis for agricultural management, facilitate data-driven decision-making processes, enable precision agriculture, drive sustainable development, reduce negative impacts on the environment, and implement the EU Green Deal-in line with internationally binding climate and environmental targets. 5 Conclusion Smart agriculture integrates advanced agricultural technologies, including AUAV, ARS, AIoT, AAI, ADT and deep learning. It can provide farmers with tools and platforms for automated irrigation, fertilization, pest control, crop pest detection, growing environment monitoring, unmanned farm management and automatic crop harvesting, so as to improve efficiency...

Gradient descent is the core and foundation of neural networks, and gradient descent optimization heuristics have greatly accelerated progress in deep learning. Although these methods are simple and effective, how they work remains unknown. Gradient descent optimization in deep learning has become a hot research topic. Some research efforts have tried to combine multiple methods to assist network training, but these methods seem to be more empirical, without theoretical guides. In this paper, a framework is proposed to illustrate the principle of combining different gradient descent optimization methods by analyzing several adaptive methods and other learning rate methods. Furthermore, inspired by the principle of warmup, CLR, and SGDR, the concept of multistage is introduced into the field of gradient descent optimization, and a gradient descent optimization strategy in deep learning model training based on multistage and method combination strategy is presented. The effectiveness of the proposed strategy is verified on the massive deep learning network training experiments.

Deep learning (DL) models have recently been widely used in UAV aerial image semantic segmentation tasks and have achieved excellent performance. However, DL models are vulnerable to adversarial examples, which bring significant security risks to safety-critical systems. Existing research mainly focuses on solving digital attacks for aerial image semantic segmentation, but adversarial patches with physical attack attributes are more threatening than digital attacks. In this article, we systematically evaluate the threat of adversarial patches on the aerial image semantic segmentation task for the first time. To defend against adversarial patch attacks and obtain accurate semantic segmentation results, we construct a novel robust feature extraction network (RFENet). Based on the characteristics of aerial images and adversarial patches, RFENet designs a limited receptive field mechanism (LRFM), a spatial semantic enhancement module (SSEM), a boundary feature perception module (BFPM) and a global correlation encoder module (GCEM), respectively, to solve adversarial patch attacks from the DL model architecture design level. We discover that semantic features, shape features and global features contained in aerial images can significantly enhance the robustness of the DL model against patch attacks. Extensive experiments on three aerial image benchmark datasets demonstrate that the proposed RFENet has strong resistance to adversarial patch attacks compared with the existing state-of-the-art methods.