Due to the complexity of background and diversity of small targets, robust detection of infrared small targets for the trajectory correction fuze has become a challenge. To solve this problem, different from the traditional method, a state-of-the-art detection method based on density-distance space is proposed to apply to the trajectory correction fuze. First, parameters of the infrared image sensor on the fuze are calculated to set the boundary limitations for the target detection method. Second, the density-distance space method is proposed to detect the candidate targets. Finally, the adaptive pixel growth (APG) algorithm is used to suppress the clutter so as to detect the real targets. Three experiments, including equivalent detection, simulation and hardware-in-loop, were implemented to verify the effectiveness of this method. Results illustrated that the infrared image sensor on the fuze has a stable field of view under rotation of the projectile, and could clearly observe the infrared small target. The proposed method has superior anti-noise, different size target detection, multi-target detection and various clutter suppression capability. Compared with six novel algorithms, our algorithm shows a perfect detection performance and acceptable time consumption.

Transformers have achieved compelling performance in vision tasks, but their substantial computational overhead remains a major obstacle to efficient real-time visual perception, especially on resource-constrained platforms such as unmanned aerial vehicles and embedded systems. Existing acceleration strategies often focus on reducing attention complexity or incorporating lightweight convolutional operations, while relatively less attention has been paid to how stage configuration, early embedding design, and potential late-stage attention redundancy jointly affect the efficiency of compact Vision Transformers. To address this issue, we empirically compare two common architectural configurations and investigate whether a three-stage design can improve inference efficiency while preserving representation quality. Based on this analysis, we propose Multi-Level Patch Embedding (MLPE), a progressive early embedding structure that goes beyond simple patchification and conventional convolutional stems. MLPE decomposes the stem-to-token transformation into preliminary downsampling, intermediate local feature refinement, and final compact embedding, helping retain and refine fine-grained local structures before generating compact token representations. Our empirical results further suggest that, within the evaluated compact architectures, allocating greater depth and concentrating attention layers in the final stage tends to improve the accuracy-efficiency balance. To reduce the resulting attention overhead, we propose Squeeze Multi-Head Self-Attention (SMSA), which learns a compact representative attention response in a squeezed feature space and fuses the re-scaled attention output with the input feature for feature recovery. Based on these designs, SFViT achieves a favorable accuracy-efficiency trade-off on ImageNet-1K and COCO, suggesting its practical potential for efficient real-time visual perception.