Unmanned Aerial Vehicles (UAVs) combined with target detection algorithms can enhance the detection of road cracks. In response to the challenges presented by complex crack shapes and textures, small sizes, and highly integrated backgrounds, our work developed a UAV-based road crack target detection algorithm using RCYOLO. RCYOLO was composed of a C2f_DySnakeConv module located in the ninth layer and a SimAM module situated above the SPPF, along with a dyhead attention detection head that integrated three types of attention mechanisms. Initially, the C2f_DySnakeConv was proposed to effectively extract tubular features of cracks. Subsequently, the SimAM addressed the issue of target-background fusion, enhancing feature recognition of the targets while suppressing background interference. Finally, the dyhead strategy incorporated three types of attention mechanisms, effectively resolving the issue of small target omissions. Our results showed that on the custom UAV dataset RCI, which included close-range and long-range images, RCYOLO outperformed the baseline network YOLOv8 by 5.9% in mAP@0.5, 6.5% in recall, and 9.8% in precision. On the public dataset DOTA, mAP@0.5 also exceeded YOLOv8 by 5.8%, indicating that RCYOLO performed well in other remote sensing image tasks, making this target detection algorithm more suitable for high-altitude photography of crack targets than other mainstream algorithms.

In this paper, an In_0.53Ga_0.47As electron–hole bilayer tunnel field-effect transistor (EHBTFET) with a dual-metal left-gate and an N⁺-pocket (DGNP-EHBTFET) is proposed and systematically studied by means of numerical simulation. Unlike traditional transverse EHBTFETs, the proposed DGNP-EHBTFET can improve device performance without sacrificing the chip density, and can simplify the manufacturing process. The introduction of the dual-metal left-gate and the N⁺-pocket can shift the point-tunneling junction and adjust the energy band and the electric field in it, aiming to substantially degrade the OFF-state current (I_OFF) and maintain good ON-state performance. Moreover, the line tunneling governed by the tunneling-gate and the right-gate can further regulate and control I_OFF. By optimizing various parameters related to the N⁺-pocket and the gate electrodes, DGNP-EHBTFET’s I_OFF is reduced by at least four orders of magnitude, it has a 75.1% decreased average subthreshold swing compared with other EHBTFETs, and it can maintain a high ON-state current. This design greatly promotes the application potential of EHBTFETs.

Natural disasters cause significant human and economic losses every year, making timely search and rescue crucial. In remote, hard-to-reach areas and during network outages after disasters, rescue missions become even more challenging. Traditional edge device solutions rely on extensive hardware resources and high power consumption to support precision and speed, which cannot meet the numerous constraints of post-disaster scenarios. Moreover, existing state-of-the-art detectors (e.g., YOLO series, RetinaNet) have chain-structured networks with narrow receptive fields and single-scale features, and cannot achieve high-speed inference through parallel computing. This paper proposes a custom convolutional neural network (CNN) solution for search and rescue, equipped with FPGA accelerators on drones. This method is autonomous, low-cost, highly accurate, and efficient for drone rescue missions. Tested on the LADD dataset, it achieved an accuracy of 87.5%, inference speed of 24.7 ms, and an efficiency ratio of 19.1. Additionally, the SRNET network incorporates an FEM structure, which expands the receptive field through a multi-scale parallel architecture, well-suited for the parallel processing capabilities of FPGAs. A reconfigurable FPGA accelerator was also developed, reusing different modules based on the instruction set. It innovatively employs kernel partitioning, significantly reducing redundant access to input feature maps and kernels, balancing parallel computation capabilities while consuming less system power. In large-scale disaster search and rescue, this approach outperforms other systems.

UAV infrared sensor technology plays an irreplaceable role in various fields. High-altitude infrared images present significant challenges for feature extraction due to their uniform texture and color, fragile and variable edge information, numerous background interference factors, and low pixel occupancy of small targets such as humans, bicycles, and diverse vehicles. In this paper, we propose a Multi-scale Dual-Branch Dynamic Feature Aggregation Network (MDDFA-Net) specifically designed to address these challenges in UAV infrared image processing. Firstly, a multi-scale dual-branch structure is employed to extract multi-level and edge feature information, which is crucial for detecting small targets in complex backgrounds. Subsequently, features at three different scales are fed into an Adaptive Feature Fusion Module for feature attention-weighted fusion, effectively filtering out background interference. Finally, the Multi-Scale Feature Enhancement and Fusion Module integrates high-level and low-level features across three scales to eliminate redundant information and enhance target detection accuracy. We conducted comprehensive experiments using the HIT-UAV dataset, which is characterized by its diversity and complexity, particularly in capturing small targets in high-altitude infrared images. Our method outperforms various state-of-the-art (SOTA) models across multiple evaluation metrics and also demonstrates strong inference speed capabilities across different devices, thereby proving the advantages of this approach in UAV infrared sensor image processing, especially for multi-scale small target detection.

In order to improve the reverse voltage capacity and low junction temperature characteristics of the traditional silicon-based Schottky diode, a Schottky diode with high reverse voltage capacity and high junction temperature was fabricated using ion implantation, NiPt60 sputtering, silicide-forming and other major technologies on an N-type silicon epitaxial layer of 10.6–11.4 μm and (2.2–2.4) × 10¹⁵ cm⁻³ doping concentration. The measurement results show that the junction temperature of the Schottky diode fabricated can reach 175 °C, that is 50 °C higher than that of the traditional one; the reverse voltage capacity V_R can reach 112 V, that is 80 V higher than that of the traditional one; the leakage current is only 2 μA and the forward conduction voltage drop is V_F = 0.71 V at forward current I_F = 3 A.