Infrared and visible images are fused to obtain a high-quality visible image with salient infrared targets that can be used in other vision tasks. The fusion strategy is key to achieving efficient image fusion. The strategies currently used in deep learning-based image fusion methods are based on manual computation and are unlearnable, limiting further improvements in infrared and visible image fusion. In this paper, FusionGRAM, a novel end-to-end infrared and visible image fusion framework, is proposed. It adaptively fuses visible detail information and infrared thermal information. The framework is fully convolutional, and the parameters of each part can be learned. The proposed method applied dense connections with an attention mechanism and gradient residual, thereby improving the network’s ability to obtain critical information during feature extraction. Further, pixel intensity loss and detail loss functions were used to train FusionGRAM. Fused features were reconstructed using four convolution layers to obtain an informative fused image. Ablation experiments and comparisons on public datasets showed that FusionGRAM outperformed current state-of-the-art fusion methods in subjective and objective evaluations.

Detecting small objects in infrared images remains a challenge because most of them lack shape and texture. In this study, we proposed an infrared small-object detection method to improve the capacity for detecting thermal objects in complex scenarios. First, a sparse-skip connection block is proposed to enhance the response of small infrared objects and suppress the background response. This block is used to construct the detection model backbone. Second, a region attention module is designed to emphasize the features of infrared small objects and suppress background regions. Finally, a batch-averaged biased classification loss function is designed to improve the accuracy of the detection model. The experimental results show that the proposed small-object detection framework significantly increases precision, recall, and F1-score, showing that, compared with the current advanced detection models for small-object detection, the proposed detection framework has better performance in infrared small-object detection under complex backgrounds. The insights gained from this study may provide new ideas for infrared small object detection and tracking.

Multimodal image fusion aims to retain valid information from different modalities, remove redundant information to highlight critical targets, and maintain rich texture details in the fused image. However, current image fusion networks only use simple convolutional layers to extract features, ignoring global dependencies and channel contexts. This paper proposes GRPAFusion, a multimodal image fusion framework based on gradient residual and pyramid attention. The framework uses multiscale gradient residual blocks to extract multiscale structural features and multigranularity detail features from the source image. The depth features from different modalities were adaptively corrected for inter-channel responses using a pyramid split attention module to generate high-quality fused images. Experimental results on public datasets indicated that GRPAFusion outperforms the current fusion methods in subjective and objective evaluations.

Velocity measurement has a high application value in hydrological monitoring and flood disaster warning. The long-distance laser Doppler water flow velocimetry technology has the advantage of strong anti-interference ability and high spatial resolution, and it can realize the high-precision measurement of water flow velocity. Because water flow has low reflectance characteristics, how to extract Doppler frequency from weak non-stationary coherent signals is a crucial problem to be solved to realize long-distance water flow velocity measurement. However, the classical method requires the time domain signal to have high stationarity and is not suitable for processing the coherent signal in the water flow velocity measurement. Aiming at this problem, we proposed a water flow velocime-try method based on adaptive Gaussian weighted integral (AGWI). First, the spectral characteristics of the coherent signal are analyzed in detail, and a statistical model of weak non-stationary signals is established. A second-order Kaiser self-multiplication window (KSMW) is designed to suppress spectral leakage for the asynchronously sam-pled data. Then, an adaptive homogenization power spectral subtraction (AHPSS) is designed to reduce system noise. Finally, the Doppler spectrum reconstruction and Doppler frequency estimation are performed using the AGWI method to obtain the Doppler frequency, which is further processed to get the water flow velocity. The exper-imental results show that the method proposed in this paper can achieve accurate and stable measurement of river surface velocity under long-distance conditions. (c) 2022 Optica Publishing Group

This paper presents a feature-level defective pixel cluster (DPC) correction method specifically designed for dual-band infrared detectors. Due to inherent manufacturing limitations, antimonide-based type-II superlattice (T2SL) focal plane arrays (FPAs) commonly suffer from DPC issues. Existing DPC correction methods often fail to preserve dual-band imaging characteristics, leading to texture distortion and incomplete structure recovery. Inspired by infrared imaging mechanisms and image inpainting theory, we propose PRISM, a patch-based reconstruction framework that leverages inter-band structure migration with hierarchical feature decomposition, decoupling dual-band images into micro-textures, edge gradients, and spatial relationships. Leveraging this multi-level feature representation, we extract prior information to guide DPC correction. The correction process is formulated as a “structure-to-pixel” optimization problem, where an improved feature-guided patch search strategy effectively combines structure completion with texture reconstruction. Experimental results demonstrate that the proposed method not only recovers image content effectively but also faithfully preserves band-specific imaging characteristics. Furthermore, to facilitate quantitative evaluation, we construct a benchmark dataset containing simulated DPCs with corresponding ground truth. Comparative experiments on both our self-constructed dataset and public multimodal benchmarks confirm that PRISM achieves higher PSNR, SSIM, VIF, and CC metrics than state-of-the-art multimodal inpainting methods.