This study based on co-occurrence analysis shearlet transform (CAST) effectively combines the latent low rank representation (LatLRR) and the regularization of zero-crossing counting in differences to fuse the heterogeneous images. First, the source images are decomposed by CAST method into base-layer and detail-layer sub-images. Secondly, for the base-layer components with larger-scale intensity variation, the LatLRR, is a valid method to extract the salient information from image sources, and can be applied to generate saliency map to implement the weighted fusion of base-layer images adaptively. Meanwhile, the regularization term of zero crossings in differences, which is a classic method of optimization, is designed as the regularization term to construct the fusion of detail-layer images. By this method, the gradient information concealed in the source images can be extracted as much as possible, then the fusion image owns more abundant edge information. Compared with other state-of-the-art algorithms on publicly available datasets, the quantitative and qualitative analysis of experimental results demonstrate that the proposed method outperformed in enhancing the contrast and achieving close fusion result.

The image decomposition strategy that extracts salient features from the source image is crucial for image fusion. To this end, we proposed a novel saliency-based decomposition strategy for infrared and visible image fusion. In particular, the latent low-rank representation (LatLRR) and rolling guidance filter (RGF) are together employed to process source images, which is called DLatLRR_RGF. In this method, the source images are first decomposed to salient components and base components based on LatLRR, and the salient components are filtered by RGF. Then, the final base components can be calculated by the difference between the source image and the processed salient components. The fusion rule based on the nuclear-norm and modified spatial frequency is used to fuse the salient components. The base components are fused by the l₂-energy minimization model. Finally, the fused image can be obtained by the fused base components and saliency detail components. Multiple groups of experiments on different pairs of infrared and visible images demonstrate that, compared with other state-of-the-art fusion algorithms, our proposed method possesses superior fusion performance from subjective and objective perspectives.

The principle of the image fusion is to integrate complementary information of the heterogeneous images to obtain a fused image that is more in line with the visual effect of the human eyes. However, most decomposition methods cannot distinguish the textures and edges in an image, which is easy to produce the halo artifacts around edges. In this paper, we proposed a novel image decomposition strategy (co-occurrence analysis shearlet transform, CAST) to preprocess the input images depending on the co-occurrence statistic information to generate the base layer and detail layer components. In order to improve the sparseness of the base layer, the classified sparse dictionary in the measurement domain is introduced to enhance the robustness of incorrect registration. As for the detail layers, the adaptive dual-channel PCNN model is adopted as the fusion rule, in which the neurons are activated by the improved spatial frequency operator (ISF), and the model uses the sum of the improved weighted Laplacian energies (IWSWL) as the adaptive linking strength. Finally, the fused image can be generated by the inverse CAST. Based on the combination of the sparseness of the classified dictionary and the visual characteristics of PCNN model, the more valuable information of the source images can be fused, so that the final fused images conform to the human visual system. Qualitative and quantitative experimental results demonstrate the superiority of the proposed method over other typical fusion techniques on the publicly available datasets.

To solve the problem that small drones in the sky are easily confused with background objects and difficult to detect, according to the characteristics of irregular movement, small size, and changeable shape of drones, using a regional target recognition algorithm, the structure characteristics of Group Convolution (GC) in Resnext50 are absorbed. The optimized GC-faster-RCNN is obtained by improving the Fast-RCNN algorithm and the following methods are performed. First, a clustering method is used to analyze the dataset, and appropriate prior bounding box types are obtained. Second, the Resnext50 is used to replace the original feature extraction network, and the improved channel attention mechanism is integrated into its network output to enhance its feature map information. Then, we calculate its effective receptive field according to the Feature Pyramid Network (FPN) structure and redesign the prior bounding box of the corresponding size to construct a multi-scale detection network for small targets. Experiments show that the algorithm has a recognition accuracy of up to 94.8% under 1080 P image quality, and a recognition speed of 8FPS, which can effectively detect the positions of 1–5 small UAVs in a picture. This method provides an effective positioning detection for low-altitude UAVs.

The light field camera records spatial and angular information of a scene within one shot, which is a unique advantage, especially for depth estimation. Both occlusion and noise make it difficult to estimate the depth of the light field. To address this problem, we propose a depth estimation method using inline occlusion handling, which uses digital refocusing to obtain refocused images and define occlusion types in scenes. An inline occlusion framework is used to construct the correlation cost. In order to remove noise, we use a variety of filtering strategies to optimize the cost. The quantitative results show that, compared with the best performance CAE, our method can reduce the MSE and bad point rate by 40.68% and 25.76%, respectively, in the challenging noise scenes of Old HCI datasets. In New HCI datasets, the reduction is 34.98% and 31.26%, respectively. The qualitative results show that we can preserve various fine structures of the real light field. Therefore, our method has significant advantages in high-noise scenes, which can better deal with the occlusion problem of depth estimation in noisy scenes.

Hyperspectral classification is a task of significant importance in the field of remote sensing image processing, with attaining high precision and rapid classification increasingly becoming a research focus. The classification accuracy depends on the degree of raw HSI feature extraction, and the use of endless classification methods has led to an increase in computational complexity. To achieve high accuracy and fast classification, this study analyzes the inherent features of HSI and proposes a novel spectral–spatial feature extraction method called window shape adaptive singular spectrum analysis (WSA-SSA) to reduce the computational complexity of feature extraction. This method combines similar pixels in the neighborhood to reconstruct every pixel in the window, and the main steps are as follows: rearranging the spectral vectors in the irregularly shaped region, constructing an extended trajectory matrix, and extracting the local spatial and spectral information while removing the noise. The results indicate that, given the small sample sizes in the Indian Pines dataset, the Pavia University dataset, and the Salinas dataset, the proposed algorithm achieves classification accuracies of 97.56%, 98.34%, and 99.77%, respectively. The classification speed is more than ten times better than that of other methods, and a classification time of only about 1–2 s is needed.