Multispectral image reconstruction, which aims to recover a three-dimensional (3D) spatial-spectral signal from a two-dimensional measurement in a spectral camera based on ghost imaging via sparsity constraint (GISC), has been attracting much attention recently. However, faced with abundant 3D spectral data, the reconstruction quality cannot meet the visual requirements. Based on the robust data processing capability of deep learning, a novel network called SSTU-Net3+ is constructed by improving U-Net3+ with a spatial-spectral transformer (SST). To enhance the feature representation of images during reconstruction, mixed pooling modules and new convolution processes are proposed to improve the performance of the encoder and decoder, with U-Net3+ as the backbone. To boost the quality of reconstructed images, with split and concatenate (Concat) operations, we construct SST modules by exploiting both spatial and spectral correlations of multispectral images to refine the spatial and spectral features. Furthermore, we employ the SST in the decoder to reconstruct the desired 3D cube. Given similar network parameters, experiments on GISC spectral imaging data show that, compared to convolutional neural network-based methods, the average peak signal-to-noise ratio of images reconstructed using SSTU-Net3+ is improved by 3%, the structural similarity is enhanced by 3%, and the spectral angle mapping is cut by 12%. Particularly, compared to differential ghost imaging and compressed sensing, the reconstruction quality of SSTU-Net3+ has been significantly improved. SSTU-Net3+ can process a large amount of 3D multispectral image data more efficiently and construct the target image more accurately than the abovementioned methods.
Roadside unit (RSU) plays an important role in communication relay and data processing in the Internet of Vehicles. The rational deployment of roadside units directly affects the operational efficiency and system robustness of the Internet of Vehicles. This article proposes an RSU deployment scheme based on vehicular transmission demand. First, the mobility model of vehicles is mined through system communication transmission requirements to determine the total delay of vehicle communication transmission tasks. The delay includes two parts, namely, the vehicle data transmission delay and the channel service delay. Second, the optimization objective of the model is established, and some parameters such as inhibition distance coefficient are defined as constraint conditions. Finally, a evolutionary feature‐based RSU deployment algorithm, ICA‐EFRDA, preprocessed by a greedy algorithm is proposed to implement RSU deployment. The deployment results show that the proposed scheme has lower delay and higher coverage time ratio than the EFRDA scheme in both urban and suburban road networks. Under limited cost constraints, the system decides to place a limited number of RSUs in low‐traffic density areas of the suburban road network to improve performance. When the traffic density is high, the increased placement of RSUs in the urban road network has a more obvious effect on improving the service capability of the system. Under different deployment cost constraints, ICA‐EFRDA outperforms the comparison schemes in terms of the contacts per trip, contact probability, and packet delivery ratio.
Aiming to address the problem of unknown dynamic target trajectory prediction and search path optimization in unmanned aerial vehicle (UAV) swarm path planning, this paper proposes a target search algorithm based on a modified target probability map (TPM). First, using the TPM, the proposed algorithm generates a high-probability distribution region of a target with directionality to fit the target trajectory and realizes the trajectory prediction of an unknown dynamic target. Then, the distributed ant colony (ACO) algorithm and the artificial potential field (APF) algorithm are combined to generate and optimize the UAV swarm search result and return path with the goal of maximizing task execution efficiency. Finally, the Monte Carlo simulation method is used to analyze the effectiveness of the proposed algorithm, and the results are evaluated from five aspects, including the number of targets captured. The simulation results show that under the condition of an unknown dynamic target trajectory, the average target captured rate and average unknown region search rate of the MTPM method were higher than that of the traditional TPM method, and the performance was improved by 14.6% and 10.7%, respectively.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.