Search

Cross-domain object detection is a key problem in the research of intelligent detection models. Different from lots of improved algorithms based on two-stage detection models, we try another way. A simple and efficient one-stage model is introduced in this paper, comprehensively considering the inference efficiency and detection precision, and expanding the scope of undertaking cross-domain object detection problems. We name this gradient reverse layer-based model YOLO-G, which greatly improves the object detection precision in cross-domain scenarios. Specifically, we add a feature alignment branch following the backbone, where the gradient reverse layer and a classifier are attached. With only a small increase in computational, the performance is higher enhanced. Experiments such as Cityscapes→Foggy Cityscapes, SIM10k→Cityscape, PASCAL VOC→Clipart, and so on, indicate that compared with most state-of-the-art (SOTA) algorithms, the proposed model achieves much better mean Average Precision (mAP). Furthermore, ablation experiments were also performed on 4 components to confirm the reliability of the model. The project is available at https://github.com/airy975924806/yolo-G. [ABSTRACT FROM AUTHOR]

The corrosion protection of deep‐sea equipment using the economical method of thermal spraying has become a research hotspot. This study investigated the corrosion behavior of high‐velocity arc‐sprayed Al–Ti–Si–RE coating in a simulated deep sea with regard to microstructures, compositions, and electrochemical measurements. Results showed that the Al–Ti–Si–RE coating has excellent corrosion resistance at high hydrostatic pressure (HP). HP plays an important role in the anodic reaction of the coating, whereas dissolved oxygen (DO) mainly affects the cathodic reaction. HP accelerates the dissolution of the coating's passive film, which results in steady corrosion. The amount of corrosion products depends on DO. The corrosion of the self‐sealing coating is inhibited in low DO when steady corrosion is achieved. [ABSTRACT FROM AUTHOR]

Phosphorescent white Light Emitting Diodes (LEDs) serving as military vehicle lighting sources have great potential for high-speed communications. Considering the light intensity, driven by LED's driving current in military vehicles, varies with the combat mission, this paper starts with analyzing the driving current's effect on phosphorescent white LED performance in two cases: LED with and without blue filter. We conclude that photoelectric characteristics and frequency response of LED without blue filter are steadier to the driving current's fluctuation. In view of good lighting and reliable communication, we argue that blue filter is not conducive to visible light communication (VLC) in the military vehicles. Comparing to phosphorescent white LED with blue filter, we observe that the one without blue filter has higher frequency severe attenuation and lower -3 dB bandwidth, which blocks its high-speed transmission application. For that, we put forward a new second-order hardware pre-equalization strategy, and the experiment results demonstrate that the -3 dB bandwidth of 1W phosphorescent white LED without blue filter can be extended from 3.8 MHz to 93 MHz, which allows 236 Mbps transmission rate with Bit Error Ratio (BER) less than 3.8×10−3 in 1.5 m by Non Return to Zero-On Off Keying(NRZ-OOK) modulation technology and meets the military intra-vehicle communication requirement in this paper. [ABSTRACT FROM AUTHOR]

The camphor sawfly, Mesoneura rufonota Rohwer, is an economically important leaf-chewing pest of the genus Cinnamomum Trew. However, little is known about the range of species that it can infest within this genus or whether larvae show preferences for particular species. We conducted preference and performance experiments under laboratory conditions to assess larval host preference and suitability as a host plant of five congeneric trees species: C. camphora (Linn) Presl, C. bodinieri Levl., C. burmanni (Nees et T. Nees) Blume, C. pauciflorum Nees, and C. micranthum (Hay.) Hay. In no-choice, two-choice and multiple-choice feeding trials, significantly higher feeding rates were found on C. camphora, followed by C. bodinieri, compared to those on the other three tree species. In two-choice behavior trials, larvae moved to occupy leaves of C. camphora faster and more frequently, followed by C. bodinieri, than when offered the other three tree species. In no-choice fitness trials, the survival of larval and pupal stage was the highest, the developmental duration of larval and pupal stage was the shortest, the pupal body weight was the heaviest and adults lived the longest on C. camphora followed by C. bodinieri, while M. rufonota did not complete development on C. burmanni, C. pauciflorum or C. micranthum. Our results demonstrate that larvae consistently prefer and perform better on C. camphora leaves, that they can utilize C. bodinieri, but it is less preferred, and that C. burmanni, C. pauciflorum, and C. micranthum appear to be unsuitable as a host for M. rufonota. [ABSTRACT FROM AUTHOR]