For drone-based detection tasks, accurately identifying small-scale targets like people, bicycles, and pedestrians remains a key challenge. In this paper, we propose DV-DETR, an improved detection model based on the Real-Time Detection Transformer (RT-DETR), specifically optimized for small target detection in high-density scenes. To achieve this, we introduce three main enhancements: (1) ResNet18 as the backbone network to improve feature extraction and reduce model complexity; (2) the integration of recalibration attention units and deformable attention mechanisms in the neck network to enhance multi-scale feature fusion and improve localization accuracy; and (3) the use of the Focaler-IoU loss function to better handle the imbalanced distribution of target scales and focus on challenging samples. Experimental results on the VisDrone2019 dataset show that DV-DETR achieves an mAP@0.5 of 50.1%, a 1.7% improvement over the baseline model, while increasing detection speed from 75 FPS to 90 FPS, meeting real-time processing requirements. These improvements not only enhance the model’s accuracy and efficiency but also provide practical significance in complex, high-density urban environments, supporting real-world applications in UAV-based surveillance and monitoring tasks.

With the rapid development of space exploration technology, the detection of extraterrestrial bodies has become increasingly important. Among these, path planning and target recognition and positioning technologies are particularly critical for applications in intelligent agents with low computational power operating in complex environments. This paper presents a novel approach to path planning on low-resolution lunar surface maps by introducing an improved A* algorithm with an adaptive heuristic function. This innovation enhances robustness in environments with limited map accuracy and enables paths that maintain maximum distance from obstacles. Additionally, we innovatively propose the Dynamic Environment Target Identification and Localization (DETIL) algorithm, which identifies unknown obstacles and employs spatiotemporal clustering to locate points of interest. Our main contributions provide valuable references for the aerospace industry, particularly in lunar exploration missions. The simulation results demonstrate that the improved A* algorithm reduces the maximum elevation difference by 55% and the maximum cumulative elevation difference by 68% compared to the traditional A* algorithm. Furthermore, the DETIL algorithm’s obstacle identification component successfully recognizes all the obstacles along the path, and its spatiotemporal clustering improves the average number of target discoveries by 152% over the conventional DBSCAN clustering approach.