{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/infrared-small-uav-target-detection-based-on","title":"Infrared Small UAV Target Detection Based on Depthwise Separable Residual Dense Network and Multiscale Feature Fusion","arxiv_id":null,"date":"2022-08-11","proceeding":"IEEE Transactions on Instrumentation and Measurement 2022 8","authors":["Houzhang Fang; Lan Ding; Liming Wang; Yi Chang; Luxin Yan; Jinhui Han"],"abstract":"Unmanned aerial vehicles (UAVs) have been widely\r\napplied in military and civilian fields, but they also pose great\r\nthreats to restricted areas, such as densely populated areas and\r\nairports. Thermal infrared (IR) imaging technology is capable\r\nof monitoring UAVs at a long range in both day and night\r\nconditions. Therefore, the anti-UAV technology based on thermal\r\nIR imaging has attracted growing attention. However, the images\r\nacquired by IR sensors often suffer from small and dim targets,\r\nas well as heavy background clutter and noise. Conventional\r\ndetection methods usually have a high false alarm rate and low\r\ndetection accuracy. This article proposes a detection method that\r\nformulates the UAV detection as predicting the residual image\r\n(i.e., background, clutter, and noise) by learning the nonlinear\r\nmapping from the input image to the residual image. The\r\nUAV target image is obtained by subtracting the residual image\r\nfrom the input IR image. The constructed end-to-end U-shaped\r\nnetwork exploits the depthwise separable residual dense blocks in\r\nthe encoder stage to extract the abundant hierarchical features.\r\nBesides, the multiscale feature fusion and representation block\r\nis introduced to fully aggregate multiscale features from the\r\nencoder layers and intermediate connection layers at the same\r\nscale, as well as the decoder layers at different scales, to better\r\nreconstruct the residual image in the decoder stage. In addition,\r\nthe global residual connection is adopted in the proposed network\r\nto provide long-distance information compensation and promote\r\ngradient backpropagation, which further improves the performance\r\nin reconstructing the image. The experimental results\r\nshow that the proposed method achieves favorable detection\r\nperformance in real-world IR images and outperforms other\r\nstate-of-the-art methods in terms of quantitative and qualitative\r\nevaluation metrics.","url_abs":"https://ieeexplore.ieee.org/abstract/document/9855493/","url_pdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=9855493","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"infrared-small-uav-target-detection-based-on","repo_url":"https://github.com/IVPLaboratory/RIPNet","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":0,"framework":"pytorch","reach":null}],"tasks":[{"task_slug":"decoder","task_name":"Decoder"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":null,"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}