Search

To address the challenges of low detection rate and high missed detection rate of military aircraft in current complex remote sensing data, and to meet the requirements of real-time detection and easy deployment of models, this article introduces DET-you only look once (YOLO), an innovative detection model. First, to tackle the issue of reduced accuracy in identifying small targets amidst intricate backgrounds, a novel feature extraction component, C2f_DEF, was devised. This module replaced all existing C2f components within YOLOv8n, thereby significantly enhancing the model's ability to cope with complicated environmental contexts. Second, to achieve the functionality of easy deployment of the model, some deep structures were simplified to make the model more lightweight. Afterward, to further improve the model's ability to handle complex backgrounds and dense environments in remote sensing images and to improve the model's detection accuracy for military aircraft, the DAT module was embedded in the model. Finally, this article also optimized the loss function and reg_max to further reduce computational costs while improving the detection accuracy of the model. To verify the effectiveness and strong universality of DET-YOLO, extensive experimental verification was conducted on three publicly available datasets, namely MAR20, NWPU VHR-10, and NEU-DET. On the MAR20 dataset, compared with other advanced models, DET-YOLO achieved the highest mAP0.5 (namely 94.7%) with only 80 training epochs while meeting lightweight and real-time requirements. While on the other two datasets, DET-YOLO also achieved the best detection performance.

Due to environmental friendliness, small size, long lifetime, and adjustable light-emitting wavelength, AlGaN-based deep ultraviolet light-emitting diodes (DUV-LEDs) have great development potential in the fields of biochemical detection, sterilization and disinfection, and non-line-of-sight communication. After decades of research, AlGaN-based DUV-LEDs have significantly improved in the direction of luminous efficiency. However, the luminous efficiency of the present AlGaN-based DUV-LEDs is still at a low level compared with that of the fully commercialised nitride blue LEDs, making it difficult to meet the market demand. This paper first introduces the present development status of AlGaN-based DUV-LEDs and the reasons for the low luminous efficiency. Then from the preparation of AlGaN-based, AlGaN doping and the design of AlGaN-based DUV-LEDs structure to elaborate the research progress in recent years in the direction of improving the luminous efficiency of AlGaN-based DUV-LEDs. Finally discusses the challenges that the AlGaN-based DUV-LEDs are presently faced with, and their future opportunities for development.

Due to the uncertain, diverse, and light-attenuating characteristics of the underwater environment, underwater images have low contrast and unclear problems. This paper proposes a histogram equalization algorithm based on optimized adaptive image quadruple segmentation and cropping (AQSCHE). Compared with the traditional histogram equalization underwater image enhancement algorithm, this algorithm introduces histogram quadruple segmentation and cropping technology. Using the exposure value and segmentation point calculation formula that optimizes the distribution range of the histogram, perform quadruple segmentation on the image to obtain a more refined histogram. The adaptive histogram clipping is realized by constructing the clipping parameter z to adjust the contrast and brightness of the image. The original image is enhanced by double equalization of the sub-histogram and the histogram of each channel. Finally, the simulation experiments verify the enhancement effect of the proposed algorithm AQSCHE on underwater images. The processed underwater image has higher contrast, is clearer and more natural in subjective evaluation, and has a better visual effect; in the image objective evaluation indicators, information entropy (Entropy), peak signal-to-noise ratio (PSNR), structural similarity index (SSIM) and universal color image quality evaluator (UCIQE), this algorithm also outperforms other common algorithms such as HE and CLAHE.

Past studies have demonstrated the positive impact of step-graded $p$ -type AlxGa1−xN/GaN superlattice (SL) electron blocking layer (EBL) structures on the efficiency performance of ultraviolet (UV) GaN-based light-emitting diodes (LEDs). However, the optimal Al-grading structure of these SL EBLs remains unclear owing to a lack of systematic investigation. The present work addresses this issue by applying EBL composed of alternating half-peak, single-peak, and double-peak $p$ -type AlxGa1−xN/GaN SL structures with varying values of $x$ ranging from 0.05 to 0.15 in 0.05 step increments. Simulation analysis is employed to obtain the internal quantum efficiency (IQE), energy band diagrams, polarization compensation factor, hole concentration, and electron concentration of GaN-based UV LEDs with three different SL-EBL structures. The results obtained at an injection current of 200 mA demonstrate that UV LEDs with double-peak SL-EBL structures provide the maximum IQE, which is ~38% greater than that of devices employing the conventional EBL in simulation experiment. This SL-EBL is demonstrated to improve the hole injection and electron overflow performance of GaN-based UV LEDs owing to the polarization charge and lattice mismatch at the p-AlxGa1−xN/GaN interfaces. The reduced Al composition on the p-GaN side reduces the potential barrier of hole injection. Moreover, the predicted increase in the IQE of GaN-based UV LEDs with the optimal SL-EBL is verified experimentally.

Rail transit is a kind of rail transit based on fixed lines, which is carried out through special vehicles, passengers, and goods. However, urban rail transit construction activities are mostly underground activities, the construction environment and construction technology are complex, and the potential disaster risks are different. At present, the traditional 3D simulation technology is lack of interactive experience, but with the development of VR technology, 3D panorama technology, as a branch of VR technology, is rapidly popularized with its interactive and immersive characteristics. VR reality technology is applied in aerospace, real estate, education, entertainment, games, and other fields. Therefore, authors propose the application of VR panoramic technology based on the Unity3D engine and HTC Vive interactive equipment in the urban rail transit vehicle engineering simulation system. Furthermore, the content construction scheme and the overall development process of the simulation system are given. The function of VR technology in the simulation system and the main technology of real-time interaction for vehicle maintenance are discussed. The methods of literature analysis, expert interviews, on-site data acquisition, and questionnaire surveys are applied. By determining the intensity of protection measures, improving the construction quality of urban rail transit, preventing and controlling risks, and making the best plan and measures from a systematic point of view. From the final results of this study, user can gradually change the traditional ideas and methods of urban rail transit disaster prevention and mitigation and implement VR technology in the urban rail transit vehicle engineering simulation system, which plays a very important role in improving the safety management efficiency of urban rail transit construction.

With the development of science and technology, Virtual Reality (VR) technology is more and more mature, and has been widely used in real life. In particular, the panorama technology in VR enables users to interact with the objects in the virtual environment in a natural way through the establishment of an artificial three-dimensional virtual environment. so as to get a good experience. Although VR technology has got some maturity, in response to the complex environment and increasing market demand, we put forward higher requirements for VR panorama technology. Panoramic mosaic technology has been studied for a long time at domestically and abroad.. Since the concept of image fusion was put forward, image fusion technology has been developed rapidly. However, the development of image mosaic technology still faces some problems to be studied and these problems were solved. Based on the existing research, present study proposes a VR panoramic image mosaic algorithm based on particle swarm optimization.. The proposed algorithm has the advantages of small computation time, fast splicing speed, high splicing quality, and high splicing confidence. This algorithm is based on particle swarm optimization, combined with the advantages of traditional mutual information algorithm, according to the characteristics of virtual reality imaging, a new model is established. After a large number of comparative experiments, it was found thatproposed method significantly improved the ability of feature region block extraction in VR panoramic image mosaic when compared with the traditional method. At the same time of improving the accuracy, the mapping time is shortened by more than 1 / 3, and the mapping efficiency is improved. Present research work is an innovation of virtual reality panoramic technology.