Complete coverage of multiple robots for a large map is an important collaborative planning task, which is widely used in disaster search and rescue, forest fire prevention, resource exploration, and other fields. It generally focuses on coverage completion with less robot (mostly drone) occupation and higher time efficiency. However, as the map becomes larger, most existing works will fail to remain near-optimal due to escalating complexity. In this letter, we formulate the problem as two levels of traveling salesman problem (TSP) to reduce the complexity, where the lower level is single-robot local coverage planning with preset TSP solutions, and the higher level is multiple TSP (mTSP) global planning executed by multiple robots. To better adapt to dynamic scenarios, we apply a distributed multi-agent reinforcement learning (MARL) framework that allows efficient online computation, and a dense segmented Siamese network (DSSN) to achieve efficient and effective solutions. We show that compared to existing advanced methods, our strategy can effectively solve the problem with coverage costs decreased by over 30 $\%$ average and the execution time reduced to seconds in large maps. Furthermore, our network DSSN achieves an additional improvement with random settings to the previous mTSP architecture. We also discuss the influence of different robot densities and separated block sizes on the results, and demonstrate the adaptability to irregular and large-scale obstacles.

The traditional template matching strategy of optical and synthetic aperture radar (SAR) is sensitive to the nonlinear transformation between two images. In some cases, the optical and SAR image pairs do not conform to the affine transformation condition. To address this issue, this study presents a novel template matching strategy which uses the One-Class Support Vector Machine (SVM) to remove outliers. First, we propose a method to construct the similarity map dataset using the SEN1-2 dataset for training the One-Class SVM. Second, a four-step strategy for optical and SAR image registration is presented in this paper. In the first step, the optical image is divided into some grids. In the second step, the strongest Harris response point is selected as the feature point in each grid. In the third step, we use Gaussian pyramid features of oriented gradients (GPOG) descriptor to calculate the similarity map in the search region. The trained One-Class SVM is used to remove outliers through similarity maps in the fourth step. Furthermore, the number of improve matches (NIM) and the rate of improve matches (RIM) are designed to measure the effect of image registration. Finally, this paper designs two experiments to prove that the proposed strategy can correctly select the matching points through similarity maps. The experimental results of the One-Class SVM in dataset show that the One-Class SVM can select the correct points in different datasets. The image registration results obtained on the second experiment show that the proposed strategy is robust to the nonlinear transformation between optical and SAR images.

Spatial Phase Prediction from intensity image without auxiliary sensors is a challenging task due to the compressed phase information and the increase in complexity with varying optical systems. Many advanced works focus primarily on the adaptability to extended light sources, while having yet to construct a feature map that is robust and sufficient enough to represent the characteristics of the aberration, leading to some problems, such as a minimum light intensity limit, a small input size, and failing to remain optimal with different phase ranges. To tackle these problems, we apply both the Fourier transform and the multi-scale wavelet transform in the feature mapping process, constructing a high-frequency concentrated aberration feature sequence (AFS). To better adapt to the learning complexity, we formulate an effective and lightweight model, named aberration sensing convolutional neural network (ASCNN), under the supervised framework that allows efficient online computation. The results demonstrate that compared to the baseline feature, the performance of the proposed AFS improves by around 30%. Besides, the model ASCNN achieves high performance in aberration retrieval, with an average RMSE better than 0.0143 λ under the initial condition of ±0.25 λ , and the SSIM exceeding 96% between the reconstructed and ground-truth phases. Furthermore, the proposed method is evaluated on both scalar diffraction and the polarized SLM system to demonstrate the robustness and the broad potential for applications.

The recognition of point patterns in spatial vector data has important applications in geographic mapping and formation recognition. However, the application of traditional methods to spatial vector data faces two difficulties. Firstly, these data are low signal-to-noise ratio data in which the point patterns are mixed with a large number of normal point clusters; thus, it is difficult to recognize point patterns from these unstructured data using traditional clustering or machine learning methods. Secondly, the lack of edge connectivity relationships in spatial vector data directly hinders the application of graph models. Few studies have systematically solved the above difficulties. In this article, we propose a geometric feature attention scheme to overcome the above challenges. We also present an implementation of the scheme based on the graph method, termed GeoFAN, to extract and classify point patterns simultaneously in spatial vector data. Firstly, the raw data are transformed into a graph structure consisting of adjacency and attribute matrices. Secondly, a geometric feature attention module is proposed to enhance the feature representation of point patterns. Finally, the recognition results of all points are output via GeoFAN. The macro precision, recall, and $F_1$ score of five simulated point pattern types with different attributes and point numbers are 92.8%, 90.3%, and 91.5%, respectively, and GeoFAN is trained with simulated data to recognize real location-based point patterns successfully. The proposed GeoFAN showed superior performance and generalization ability in point pattern recognition.

A novel approach for laser coherent beam combination (CBC) utilizing the self-imaging effect in plasma waveguides is presented in this study, which enables the transmission of ultrashort laser pulses at intensities above the bulk damage threshold of conventional solid optical waveguides. The feasibility of self-imaging-based CBC in plasma waveguides was simulated and verified, demonstrating favorable combining efficiency and beam quality. This work explores the adaptive tuning of waveguide length via dynamic adjustment of plasma density, addressing the critical issue of fabrication tolerances in traditional waveguide systems. With CBC via plasma waveguide, this study offers support for the development of robust, high-power laser systems with enhanced beam quality and operational stability.

Spectral super-resolution (SSR) from RGB response can achieve flexible applications and high resolution (both spatial and spectral) under low-cost conditions, which is beneficial in various tasks like surface coating, target recognition, and vegetation monitoring. However, most existing SSR works are based on the overall features of the image to pursue lower average errors, while such averaging effect could weaken the improvement of higher reconstruction accuracy for individual pixels, and also leads to deeper and more complex models. To address these problems, we propose a two-stage method focused on the local phase and color similarity from a pixel perspective. In the first stage, a selection strategy named Tristimulus Color Congruency (TCC) is introduced to extract the color similarity of optical images and guide the selection of pixels to enrich the training process, which is very sparse compared to the whole image (less than 1%). In the second stage, the Spectral Extension Network (SEN) with several easy-to-deploy blocks is constructed to complete an effective and efficient single-pixel spectrum prediction. Single pixel computation time of our model is compressed to the order of μ s, and the method is evaluated in three public benchmarks (Munsell-1269, CAVE-real, and AeroRIT). In the hyperspectral image reconstruction task, our method improves the reconstruction performance by over 40% overall test compared to the baseline and shows consistency in different environments. Besides, as pixel-based, our method can also serve as a plug-and-play model to improve the performance on any region of any size or discrete pixels of interest. The reference code is available at https://github.com/Eahein-H/TCC-SEN .