Search

Unmanned aerial vehicle target tracking is a complex task that encounters challenges in scenarios involving limited computing resources, real-time requirements, and target confusion. This research builds on previous work and addresses challenges by proposing a grid-based beetle antennae search algorithm and designing a lightweight multi-target detection and positioning method, which integrates interference-sensing mechanisms and depth information. First, the grid-based beetle antennae search algorithm’s rapid convergence advantage is combined with a secondary search and rollback mechanism, enhancing its search efficiency and ability to escape local extreme areas. Then, the You Only Look Once (version 8) model is employed for target detection, while corner detection, feature point extraction, and dictionary matching introduce a confusion-aware mechanism. This mechanism effectively distinguishes potentially confusing targets within the field of view, enhancing the system’s robustness. Finally, the depth-based localization of the target is performed. To verify the performance of the proposed approach, a series of experiments were conducted on the grid-based beetle antennae search algorithm. Comparisons with four mainstream intelligent search algorithms are provided, with the results showing that the grid-based beetle antennae search algorithm excels in the number of iterations to convergence, path length, and convergence speed. When the algorithm faces non-local extreme-value-area environments, the speed is increased by more than 89%. In comparison with previous work, the algorithm speed is increased by more than 233%. Performance tests on the confusion-aware mechanism by using a self-made interference dataset demonstrate the model’s high discriminative ability. The results also indicate that the model meets the real-time requirements.

Hyperspectral images (HSIs) possess an inherent three-order structure, prompting increased interest in extracting 3D features. Tensor analysis and low-rank representations, notably truncated higher-order SVD (T-HOSVD), have gained prominence for this purpose. However, determining the optimal order and addressing sensitivity to changes in data distribution remain challenging. To tackle these issues, this paper introduces an unsupervised Superpixelwise Multiscale Adaptive T-HOSVD (SmaT-HOSVD) method. Leveraging superpixel segmentation, the algorithm identifies homogeneous regions, facilitating the extraction of local features to enhance spatial contextual information within the image. Subsequently, T-HOSVD is adaptively applied to the obtained superpixel blocks for feature extraction and fusion across different scales. SmaT-HOSVD harnesses superpixel blocks and low-rank representations to extract 3D features, effectively capturing both spectral and spatial information of HSIs. By integrating optimal-rank estimation and multiscale fusion strategies, it acquires more comprehensive low-rank information and mitigates sensitivity to data variations. Notably, when trained on subsets comprising 2%, 1%, and 1% of the Indian Pines, University of Pavia, and Salinas datasets, respectively, SmaT-HOSVD achieves impressive overall accuracies of 93.31%, 97.21%, and 99.25%, while maintaining excellent efficiency. Future research will explore SmaT-HOSVD’s applicability in deep-sea HSI classification and pursue additional avenues for advancing the field.