Bo Li

Due to the refraction, absorption, and scattering effects of light propagation in underwater environments, underwater images often suffer from color distortion, blurring, and reduced visibility. To address these challenges, frequency‐based approaches have demonstrated significant potential for improving underwater image enhancement performance. Our analysis suggests complementary roles for spatially localized wavelet subbands and Fourier amplitude–phase representations in modeling illumination and structural information. Motivated by the complementary characteristics of these two frequency‐domain representations, we propose a Wavelet Spatial Fusion Fourier‐Aware Network (WSF‐FAN) for underwater image enhancement. The proposed network adopts an encoder–latent space–decoder architecture and consists of three key components: A Fourier‐Aware Semantic Modulator (FASM), a Wavelet Low‐Frequency Spatial Attention Module (WLFSAM), and a Wavelet‐Guided Aggregation Block (WGAB). Specifically, FASM is deployed in the latent space to enhance local texture details and alleviate image blurring. WLFSAM is embedded into both the encoder and decoder to strengthen the network's perception of global illumination information. Meanwhile, WGAB is incorporated into the skip connections to adaptively emphasize informative channel features, thereby improving feature fusion quality and detail restoration capability. Extensive experiments are conducted on benchmark underwater image datasets. The results, supported by comprehensive ablation studies and comparative analyses, demonstrate competitive enhancement performance and support the contributions of joint wavelet–Fourier modeling under the evaluated settings.

Transmission lines (TLs) are the backbone of the power system, enabling efficient and reliable transmission of power over long distances. Fault diagnosis is the key technology of maintenance. Currently, there are several computer vision-based fault detection methods for power equipment. However, valuable fault samples from complex environments are difficult to collect, and many device faults exist in the form of small targets. Inspired by sketch-based retrieval, we first propose a method for TL fault diagnosis (TLFD) based on hand-drawn sketches. First, to address the problem of incomplete feature extraction, a densely connected multiscale feature network is proposed. Second, to address the problem of insufficient use of prior knowledge, a domain-aware dual attention network is proposed. Finally, a similarity loss function is constructed to supervise the similarity between all images, which preserves the prior knowledge in the pretrained model. To study the instance-level relationship between hand-drawn sketches and real images, we created the “TLFD” data. Experiments demonstrate that we provide a new idea for power system fault diagnosis, and the model significantly outperforms many state-of-the-art models designed for fault diagnosis.

To improve the positioning accuracy and efficiency of indoor visible light communication, a new indoor visible light positioning method based on the Chaotic Seagull Optimization Algorithm (CSOA) is proposed in this paper. Firstly, this paper analyzes the advantages and problems of indoor positioning technology based on visible light communication and introduces the current research status of visible light positioning at home and abroad. Secondly, the traditional seagull algorithm is improved by introducing chaotic mapping, improving control parameters, and Logistic chaos perturbation, to improve the speed and accuracy of indoor visible light communication. Moreover, compared with other intelligent algorithms, experimental data show that the visible light positioning accuracy of the improved Seagull algorithm proposed in this paper has been improved. Positioning time will also be reduced. Finally, this paper also summarizes the simulation results. In the future, the parameter and positioning models will be further optimized to improve the stability of visible light positioning.

—to address this problem of the efficiency in a V-BLAST system, an Innovative Power-restricted Allocation Algorithm in the V-BLAST System with the Adaptive Modulation using the Low Complexity was proposed in this paper. The presented method can make the data rate maximizing under the satisfaction of the peak power constraint on every antenna and the average total power limit. The simulation was implemented, and it can be found from results that the presented algorithm has the same optimal power allocation in contrast to an incremental method while its computing complexity decreases obviously; furthermore, under the same target BER, to limit the peak power for each antenna can decline the bit rate.

In low-light conditions, the detection scene can be harsh, some fundamental image features of the target to be lost, which can result in the disappearance of essential visual characteristics of the object to be detected. They have failed to balance the connection between the low-level semantic information of low-light images and normal images. This article proposes an algorithm for weak-supervised and adaptive object detection in the low-light environment for YOLOV7 (WSA-YOLO) that utilizes adaptive enhancement to effectively improve object detection capability in low-light environments, addressing this practical issue. The proposed decomposition network decomposes the image into reflectance and illumination maps, which are then enhanced separately. The proposed adaptive residual feature block (ARFB) effectively utilizes the feature correlation between low-light and normal-light images and shares the weights between them to improve parameter reuse capability during parameter prediction using the parameter prediction block. The proposed adaptive adjustment block and consistency loss function are used together to enhance the brightness and suppress noise. Finally, the you only look once (YOLO) framework is utilized for object classification, regression, and prediction. Using the metric mean average precision (mAP) for evaluation on the recognized datasets, the proposed WSA-YOLO has a performance improvement of about 8% in peak signal-to-noise ratio (PSNR), structural similarity index, and natural image quality evaluator (NIQE). And the increase in mAP is about 9%.