低空底层视觉综述

Research Review on Low-Altitude Low-Level Vision

摘要

底层视觉技术对提升无人机（UAV）在复杂环境下的感知能力至关重要。然而，低空场景特有的运动模糊、气象扰动及光照不足等耦合退化问题，叠加UAV平台的算力约束与低空复杂物理环境，严重制约了现有算法的鲁棒性与边缘端实时性。对此，文章系统综述了低空底层视觉领域的研究进展，聚焦退化恢复、信息增强与质量评估三大核心方向，不仅深入剖析了超分辨率、恶劣天气退化去除、低光增强及多源融合等前沿方法的技术特点与应用价值，还系统梳理了现有的量化评估体系，并进一步指出未来需重点突破多模态协同、无监督/自监督学习等方向，以推动低空智能感知技术的持续演进。

Abstract

Low-level visual technology is crucial for enhancing the perception capability of unmanned aerial vehicles (UAV) in complex environments. However, the coupled degradation problems unique to low-altitude scenarios, such as motion blur, meteorological disturbances, and insufficient illumination, combined with the computational constraints of UAV platforms and the complex physical environment at low altitudes, severely restrict the robustness and edge-side real-time performance of existing algorithms. To address this, this paper systematically reviews the research progress in the low-altitude low-level vision field, focusing on three core directions: degradation recovery, information enhancement, and quality assessment. This paper not only deeply analyzes the technical characteristics and application value of cutting-edge methods such as super-resolution, degradation removal under adverse weather conditions, low-light enhancement, and multi-source fusion, but also systematically summarizes the existing quantitative evaluation systems. Furthermore, this paper points out that future efforts should focus on key breakthroughs in multimodal collaboration, unsupervised/self-supervised learning, etc., to drive the continuous advancement of low-altitude intelligent perception technology.

关键词

低空底层视觉 / UAV感知 / 图像恢复 / 图像增强 / 多源融合

Key words

low-altitude low-level vision / UAV perception / image restoration / image enhancement / multi-source fusion

参考文献

[1] 董华军, 王宇栖 . 基于改进YOLOv5s的无人机小目标检测算法研究[J]. 华东交通大学学报, 2024, 41(4): 118-126.

[2] DONG H J, WANG Y Q . Research on UAV small target detection algorithm based on improved YOLOv5s[J]. Journal of East China Jiaotong University, 2024, 41(4): 118-126.

[3] 孙一铭, 赵柯嘉, 王硕, 等 . 面向无人机的低空视觉数据集研究综述[J]. 数据采集与处理, 2025, 40(2): 274-302.

[4] SUN Y M, ZHAO K J, WANG S, et al. Research review on low—altitude visual datasets for unmanned aerial vehicles[J]. Journal of Data Acquisition & Processing, 2025, 40(2): 274-302.

[5] LIANG J Y, CAO J Z, SUN G L, et al. SwinIR: image restoration using Swin transformer[C]// 2021 IEEE/CVF International Conference on Computer Vision Workshops (ICCVW), October 11—17, 2021, Montreal, BC, Canada. New York: IEEE, 2021: 1833-1844.

[6] SHANG W, REN D W, ZHANG W Y, et al. Arbitrary—scale video super—resolution with structural and textural priors[C]//LEONARDIS A, RICCI E, ROTH S, et al., eds. Computer Vision—ECCV 2024. Cham: Springer, 2025: 73-90.

[7] 李志朋, 赵长明, 张海洋, 等 . 多尺度融合超分辨率算法在无人机探测中的应用[J]. 应用光学, 2021, 42(3): 462-473.

[8] LI Z P, ZHAO C M, ZHANG H Y, et al. Application of multi—scale fusion super—resolution algorithm in UAV detection[J]. Journal of Applied Optics, 2021, 42(3): 462-473.

[9] LI R, ZHAO X W . LSwinSR: UAV imagery super—resolution based on linear Swin transformer[J]. IEEE Transactions on Geoscience and Remote Sensing, 2024, 62: 5641013.

[10] LIU J, TANG J, WU G S . Residual feature distillation network for lightweight image super—resolution[C]//BARTOLI A, FUSIELLO A, eds. Computer Vision—ECCV 2020 Workshops. Cham: Springer, 2020: 41-55.

[11] CHEN Z Y, WANG Y C, YANG Y, et al. PSD: principled synthetic—to—real dehazing guided by physical priors[C]// 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), June 20—25, 2021, Nashville, TN, USA. New York: IEEE, 2021: 7176-7185.

[12] WU R Q, DUAN Z P, GUO C L, et al. RIDCP: revitalizing real image dehazing via high—quality codebook priors[C]// 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), June 17—24, 2023, Vancouver, BC, Canada. New York: IEEE, 2023: 22282-22291.

[13] XU J Q, HU X W, ZHU L, et al. Video dehazing via a multi—range temporal alignment network with physical prior[C]// 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), June 17—24, 2023, Vancouver, BC, Canada. New York: IEEE, 2023: 18053-18062.

[14] FAN J K, WANG K, YAN Z Q, et al. Depth—centric dehazing and depth estimation from real—world hazy driving video[J]. Proceedings of the AAAI Conference on Artificial Intelligence, 2025, 39(3): 2852-2860.

[15] SU J M, LIU L L, WAN Q T, et al. Dehazing research on brightness equalization model of drone image[J]. The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, 2020, XLII—3/W10: 1289-1294.

[16] LI Z L . A transmittance optimization—based framework for image dehazing on multi—rotor drones imaging[C]// 2022 14th International Conference on Machine Learning and Computing (ICMLC), Guangzhou, China. New York: ACM, 2022: 550-554.

[17] QIU Z F, GONG T Y, LIANG Z C, et al. Perception—oriented UAV image dehazing based on super—pixel scene prior[J]. IEEE Transactions on Geoscience and Remote Sensing, 2024, 62: 5913519.

[18] REN D W, ZUO W M, HU Q H, et al. Progressive image deraining networks: a better and simpler baseline[C]// 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), June 15—20, 2019, Long Beach, CA, USA. New York: IEEE, 2019: 3932-3941.

[19] YE Y T, CHANG Y, ZHOU H Y, et al. Closing the loop: joint rain generation and removal via disentangled image translation[C]// 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), June 20—25, 2021, Nashville, TN, USA. New York: IEEE, 2021: 2053-2062.

[20] YAN W D, TAN R T, YANG W H, et al. Self—aligned video deraining with transmission—depth consistency[C]// 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), June 20—25, 2021, Nashville, TN, USA. New York: IEEE, 2021: 11961-11971.

[21] HUANG H B, YU A J, HE R . Memory oriented transfer learning for semi—supervised image deraining[C]// 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), June 20—25, 2021, Nashville, TN, USA. New York: IEEE, 2021: 7728-7737.

[22] DONG G L, ZHENG T H, CAO Y, et al. Channel consistency prior and self—reconstruction strategy based unsupervised image deraining[C]// 2025 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), June 10—17, 2025, Nashville, TN, USA. New York: IEEE, 2025: 7469-7479.

[23] XI Y, JIA W J, MIAO Q G, et al. CoDerainNet: collaborative deraining network for drone—view object detection in rainy weather conditions[J]. Remote Sensing, 2023, 15(6): 1487.

[24] YAO Y J, SHI Z M, HU H W, et al. GSDerainNet: a deep network architecture based on a Gaussian Shannon filter for single image deraining[J]. Remote Sensing, 2023, 15(19): 4825.

[25] YADAV S, MEHRA A, ROHMETRA H, et al. DerainGAN: single image deraining using Wasserstein GAN[J]. Multimedia Tools and Applications, 2021, 80(30): 36491-36507.

[26] GUO X, WANG X, FU X Y, et al. Deep unfolding network for image desnowing with snow shape prior[J]. IEEE Transactions on Circuits and Systems for Video Technology, 2025, 35(5): 4740-4752.

[27] WANG Y L, MA C, LIU J Z . SmartAssign: learning a smart knowledge assignment strategy for deraining and desnowing[C]// 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), June 17—24, 2023, Vancouver, BC, Canada. New York: IEEE, 2023: 3677-3686.

[28] LI Y X, DAI L . Video desnower: an adaptive feature fusion understanding video desnowing model with deformable convolution and KNN point cloud transformer[J]. IEEE Access, 2024, 12: 104354-104366.

[29] SALEEM A, AWAD A, MAZEN A, et al. Estimating snow coverage percentage on solar panels using drone imagery and machine learning for enhanced energy efficiency[J]. Energies, 2025, 18(7): 1729.

[30] BELMONTE A, SANKEY T, BIEDERMAN J, et al. UAV—based estimate of snow cover dynamics: optimizing semi—arid forest structure for snow persistence[J]. Remote Sensing, 2021, 13(5): 1036.

[31] XUE T H, ZHOU G, HE R L, et al. RVDNet: a two—stage network for real—world video desnowing with domain adaptation[C]// ICASSP 2024—2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), April 14—19, 2024, Seoul, Korea, Republic of. New York: IEEE, 2024: 3305-3309.

[32] GUO Y, GAO Y, LU Y X, et al. OneRestore: a universal restoration framework for composite degradation[C]// Computer Vision—ECCV 2024. Cham: Springer Nature Switzerland, 2024: 255-272.

[33] ZHANG K H, LUO W H, ZHONG Y R, et al. Deblurring by realistic blurring[C]// 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), June 13—19, 2020, Seattle, WA, USA. New York: IEEE, 2020: 2734-2743.

[34] CAO M D, FAN Y B, ZHANG Y, et al. VDTR: video deblurring with transformer[J]. IEEE Transactions on Circuits and Systems for Video Technology, 2023, 33(1): 160-171.

[35] ZHANG X Q, JIANG R H, WANG T, et al. Recursive neural network for video deblurring[J]. IEEE Transactions on Circuits and Systems for Video Technology, 2021, 31(8): 3025-3036.

[36] TRUONG N Q, LEE Y W, OWAIS M, et al. SlimDeblurGAN—based motion deblurring and marker detection for autonomous drone landing[J]. Sensors, 2020, 20(14): 3918.

[37] LIU Y, WANG J W, QIU T T, et al. An adaptive deblurring vehicle detection method for high—speed moving drones: resistance to shake[J]. Entropy, 2021, 23(10): 1358.

[38] XIAO Y Z, ZHANG J D, CHEN W, et al. SR—DeblurUGAN: an end—to—end super—resolution and deblurring model with high performance[J]. Drones, 2022, 6(7): 162.

[39] ZAMIR S W, ARORA A, KHAN S, et al. Restormer: efficient transformer for high—resolution image restoration[C]// 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), June 18—24, 2022, New Orleans, LA, USA. New York: IEEE, 2022: 5718-5729.

[40] TIAN Y F, AHMAD R B, ABDULLAH N A B . Accurate and efficient insulator maintenance: a DETR algorithm for drone imagery[J]. PLoS One, 2025, 20(2): e0318225.

[41] YANG P, WEN C W, GENG H L, et al. Intelligent fault diagnosis method for blade damage of quad—rotor UAV based on stacked pruning sparse denoising autoencoder and convolutional neural network[J]. Machines, 2021, 9(12): 360.

[42] JIN S, BAE Y, LEE S . ID—SAbRUNet: deep neural network for disturbance suppression of drone ISAR images[J]. IEEE Sensors Journal, 2024, 24(9): 15551-15565.

[43] AZMAT U, ALOTAIBI S S, ABDELHAQ M, et al. Aerial insights: deep learning—based human action recognition in drone imagery[J]. IEEE Access, 2023, 11: 83946-83961.

[44] ZHANG X, WU X L . Attention—guided image compression by deep reconstruction of compressive sensed saliency skeleton[C]// 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), June 20—25, 2021, Nashville, TN, USA. New York: IEEE, 2021: 13349-13359.

[45] YANG F, HERRANZ L, CHENG Y M, et al. Slimmable compressive autoencoders for practical neural image compression[C]// 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), June 20—25, 2021, Nashville, TN, USA. New York: IEEE, 2021: 4996-5005.

[46] YANG R, MENTZER F, VAN GOOL L, et al. Learning for video compression with hierarchical quality and recurrent enhancement[C]// 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), June 13—19, 2020, Seattle, WA, USA. New York: IEEE, 2020: 6627-6636.

[47] 叶枫, 董凡可, 贾川民 . 端到端智能视频压缩技术及其在无人机中的应用[J]. 数据采集与处理, 2025, 40(2): 303-319.

[48] YE F, DONG F K, JIA C M . End—to—end video compression technology and its application in unmanned aerial vehicles[J]. Journal of Data Acquisition and Processing, 2025, 40(2): 303-319.

[49] SMIRNOV K, MOZHAEVA A . Analysis of video data compression algorithms and optimized for use in remotely controlled drones[J]. Synchroinfo Journal, 2023, 9(2): 9-16.

[50] XU D, CHELLAPPA R, VAN GOOL L, et al. Guest editorial: special issue on deep learning for video analysis and compression[J]. International Journal of Computer Vision, 2021, 129(12): 3171-3173.

[51] ZAMIR S W, ARORA A, KHAN S, et al. Learning enriched features for real image restoration and enhancement[C]//VEDALDI A, BISCHOF H, BROX T, et al., eds. Computer Vision—ECCV 2020. Cham: Springer, 2020: 492-511.

[52] XU X G, WANG R X, FU C W, et al. SNR—aware low—light image enhancement[C]// 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), June 18—24, 2022, New Orleans, LA, USA. New York: IEEE, 2022: 17693-17703.

[53] SHE D Y, XU K . An image—to—video model for real—time video enhancement[C]// Proceedings of the 30th ACM International Conference on Multimedia, Lisboa, Portugal. New York: ACM, 2022: 1837-1846.

[54] REALE M J, MURPHY D P, CORNACCHIA M, et al. Video modification in drone and satellite imagery[C]// Disruptive Technologies in Information Sciences VIII, April 21—26, 2024. National Harbor, USA: SPIE, 2024: 41.

[55] SUN Y, ZHI X Y, HAN H W, et al. Enhancing UAV detection in surveillance camera videos through spatiotemporal information and optical flow[J]. Sensors, 2023, 23(13): 6037.

[56] MA J Y, YU W, LIANG P W, et al. FusionGAN: a generative adversarial network for infrared and visible image fusion[J]. Information Fusion, 2019, 48: 11-26.

[57] SUN Y M, CAO B, ZHU P F, et al. Dynamic brightness adaptation for robust multi—modal image fusion[J]. arXiv e—Prints, 2024: arXiv:2411.04697.

[58] GONG M Q, ZHANG H, YI X P, et al. TemCoCo: temporally consistent multi—modal video fusion with visual—semantic collaboration[J]. arXiv e—Prints, 2025: arXiv:2508.17817.

[59] LIU G C, LI C, ZHANG S H, et al. VL—MFL: UAV visual localization based on multisource image feature learning[J]. IEEE Transactions on Geoscience and Remote Sensing, 2024, 62: 5618612.

[60] LI Q P, ZHANG Y X, FANG L Y, et al. DREB—Net: dual—stream restoration embedding blur—feature fusion network for high—mobility UAV object detection[J]. arXiv e—Prints, 2024: arXiv:2410.17822.

[61] SUN Y M, LI X, ZHU P F, et al. Task—gated multi—expert collaboration network for degraded multi—modal image fusion[C]// ICML'25: Proceedings of the 42nd International Conference on Machine Learning, July 19, 2025, Vancouver, Canada. Massachusetts, USA: MIT Press, 2025: 57571-57586.

[62] GOLESTANEH S A, DADSETAN S, KITANI K M . No—reference image quality assessment via transformers, relative ranking, and self—consistency[C]// 2022 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), January 3—8, 2022, Waikoloa, HI, USA. New York: IEEE, 2022: 3989-3999.

[63] YANG S D, WU T H, SHI S W, et al. MANIQA: multi—dimension attention network for no—reference image quality assessment[C]// 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), June 19—20, 2022, New Orleans, LA, USA. New York: IEEE, 2022: 1190-1199.

基金资助

国家自然科学基金重点项目(62436002)

国家自然科学基金青年项目(62506073)

国家重点研发计划雄安新区科技创新专项(2025XAGG0039)

天津市杰出青年科学基金项目(23JCJQJC00270)

中国博士后科学基金国家资助博士后研究人员计划(GZB20250395)