FAIR1M: A Benchmark Dataset for Fine-grained Object Recognition in High-Resolution Remote Sensing Imagery
Paper • 2103.05569 • Published
| Domain | Dataset | Task | Raw Label | Path | Source | Modality | ☁️ Included | 🖥️ Server |
|---|---|---|---|---|---|---|---|---|
| 1-1 | AgroVG | Agricultural Visual Grounding 农业目标视觉定位 |
Box+Referring | ObjEarth-Data/AgroVG/t1 |
here | RGB | ✅ | ⬇️☑️ |
| 1-1 | AgroMind | agricultural remote sensing benchmark(13 tasks) 农业遥感VQA |
VQA | ObjEarth-Data/AgroMind/ |
here | RGB | ✅ | ⬇️☑️ |
| 1-1 | gRef-CW | Generalised Visual Grounding in Agriculture 农业目标视觉定位 |
Box+Referring | ObjEarth-Data/gRef-CW/ - ObjEarth-Bench/gRef-CW/ |
here | RGB | ✅ | ⬇️ |
| 1-1 | DroneBird | Bird Counting and Tracking 鸟群计数、跟踪、密度图估计 |
Point | ObjEarth-Data/DroneBird/ |
here | Video | ✅ | ⬇️☑️ |
| 1-1 | AnimalDrone | Video Animal Counting 视频动物计数 |
Point | ObjEarth-Data/AnimalDrone/ |
here | Video | ✅ | ⬇️☑️ |
| 1-1 | TinyTrees | Tree Counting 树木计数 |
Point | ObjEarth-Data/tinytrees/ |
here | Optical satellite | ✅ | ⬇️☑️ |
| 1-1 | MCPC | Multi-Category Crop Plant Counting 多类作物植物计数 |
Point | ObjEarth-Data/MCPC |
here | RGB | ✅ | ⬇️☑️ |
| 1-1 | RFCP | Rape Flower Cluster Counting 油菜花簇计数 |
Point | ObjEarth-Data/RFCP |
here | RGB | ✅ | ⬇️☑️ |
| 1-2 | STN-PLAD | Power Line Assets Detection 输电线路资产检测 |
Box | ObjEarth-Data/STN-PLAD |
here | RGB | ✅ | ⬇️☑️ |
| 1-2 | InsPLAD | Power Line Asset Inspection 输电线路资产检测 |
Box | ObjEarth-Data/InsPLAD |
here | RGB | ✅ | ⬇️☑️ |
| 1-2 | WTBD | Wind Turbine Blade Detection 风力涡轮机叶片缺陷检测 |
Box | ObjEarth-Data/WTBD |
here | RGB | ✅ | ⬇️☑️ |
| 1-2 | TTPLA | Detection and Segmentation of Transmission Towers and Power Lines 输电塔和电力线的检测分割 |
Box+Mask | ObjEarth-Data/TTPLA |
here | RGB | ✅ | ⬇️☑️ |
| 1-3 | UAVConstructionSite | UAV Construction Site Image Captioning 建筑施工现场图像描述 |
Caption | ObjEarth-Data/UAVConstructionSite |
here | RGB | ✅ | ⬇️☑️ |
| 1-3 | ELDOR | Illegal Gold Mining 亚马逊雨林非法采矿检测 |
Mask | ObjEarth-Data/ELDOR |
here | RGB | ✅ | ⬇️☑️ |
| 1-3 | CIS | Construction Instance Segmentation 建筑施工场景实例分割 |
Box+Mask | ObjEarth-Data/CIS |
here | RGB | ✅ | ⬇️☑️ |
| 1-4 | FLAME2-DT | Wildland Fire Detection 森林野火火灾检测 |
Box | ObjEarth-Data/FLAME2-DT/ |
here | RGB+Infrared | ✅ | ⬇️☑️ |
| 1-4 | SARD | Person Detection in Search and Rescue Scenarios in Drone 无人机搜救人员检测 |
Box | ObjEarth-Data/SARD-fixed |
here | RGB | ✅ | ⬇️☑️ |
| 1-4 | RescueNet | Post Disaster UAV Dataset for Semantic Segmentation 灾后无人机语义分割 |
Mask | ObjEarth-Data/RescueNet |
here | RGB | ✅ | ⬇️☑️ |
| 1-4 | TinyPerson | Person Detection for Maritime Rescue 人员检测与海事搜救 |
Box | ObjEarth-Data/TinyPerson |
here | RGB | ✅ | ⬇️☑️ |
| 1-4 | FASDD_UAV | Flame and Smoke Detection 火焰和烟雾检测 |
Box | ObjEarth-Data/FASDD_UAV |
here | RGB | ✅ | ⬇️☑️ |
| 1-5 | UAV-SEG-2025 | UAV Pavement Distress Segmentation 路面损坏检测 |
Box+Mask | ObjEarth-Data/UAV-SEG-2025 |
here | RGB | ✅ | ⬇️☑️ |
| 1-5 | UAVVaste | UAV Trash and Litter Detection 垃圾检测 |
Box | ObjEarth-Data/UAVVaste/ |
here | RGB | ✅ | ⬇️☑️ |
| 1-5 | TriModalDet | Tri-modal Object Detection 基于RGB-热成像-事件相机的三模态目标检测 |
Box | ObjEarth-Data/TriModalDet/ |
here | RGB+Thermal+Event camera | ✅ | ⬇️☑️ |
| 1-5 | CARPK | Drone-based Object Counting (car parking lot dataset) 停车场目标计数 |
Box | ObjEarth-Data/CARPK |
here | RGB | ✅ | ⬇️☑️ |
| 1-5 | PUCPR+ | 10th-floor-view Object Counting (car parking lot dataset) 停车场目标计数 |
Box | ObjEarth-Data/CARPK |
here | RGB | ✅ | ⬇️☑️ |
| 1-6 | AFO | Aerial Floating Object Detection (Maritime Search and Rescue) 水中漂浮物检测与海事搜救 |
Box | ObjEarth-Data/AFO |
here | RGB | ✅ | ⬇️☑️ |
| 1-6 | FloPWD | Floating Plastic Waste Segmentation 漂浮塑料垃圾分割 |
Mask | ObjEarth-Data/FloPWD |
here | RGB | ✅ | ⬇️☑️ |
| 1-6 | SeaDronesSee | Human Detection (Maritime Search and Rescue) 人员检测跟踪与海事搜救 |
Box | ObjEarth-Data/SeaDronesSee |
here | RGB+Video | ✅ | ⬇️☑️ |
| 1-6 | AVD | Aerial Vessels Detection 航拍船舶目标检测 |
Box | ObjEarth-Data/AVD |
here | RGB | ✅ | ⬇️☑️ |
| 1-1 | UAVPorpoises | Drone-based Porpoise Tracking 海洋哺乳动物(海豚)跟踪 |
Box | ObjEarth-Data/UAVPorpoises |
here | Video | ✅ | ⬇️☑️ |
| 2-1 | VHRShips | Ship Detection 船舶检测、定位、识别与分类 |
Box | ObjEarth-Data/VHRShips |
here | Optical satellite | ✅ | ⬇️☑️ |
| 2-1 | HRSID | Ship Detection 高分辨率SAR图像船舶目标检测 |
Box+Mask | ObjEarth-Data/HRSID |
here | SAR | ✅ | ⬇️☑️ |
| 2-1 | MAR20 | Military Aircraft Detection 遥感图像军用飞机目标检测 |
Box | ObjEarth-Data/MAR20 |
here | Optical satellite | ✅ | ⬇️☑️ |
| 2-1 | FAIR1M (1,732 images been released) | Object Detection in High-Resolution Satellite Images 高分辨率卫星图像目标检测 |
Box | ObjEarth-Data/FAIR1M |
here | Optical satellite | ✅ | ⬇️☑️ |
| 2-1 | ShipRSImagaeNet | Ship Detection 高分辨率遥感图像船舶目标检测 |
Box | ObjEarth-Data/ShipRSImagaeNet |
here | Optical satellite | ✅ | ⬇️☑️ |
| 2-1 | SADD | SAR Aircraft Detection SAR图像飞机检测 |
Box | ObjEarth-Data/SADD |
here | SAR | ✅ | ⬇️☑️ |
| 2-1 | LEVIR-Ship | Tiny Ship Detection 微型船舶目标检测 |
Box | ObjEarth-Data/LEVIR-Ship |
here | Optical satellite | ✅ | ⬇️☑️ |
| 2-1 | xView | Object Detection for Overhead Imagery 遥感目标检测 |
Box | ObjEarth-Data/xView |
here | Optical satellite | ✅ | ⬇️ |
| 2-1 | DOTA | Object Detection in Aerial Images 航拍目标检测 |
Box | ObjEarth-Data/DOTA |
here | RGB | ✅ | * |
| 2-1 | fMoW-rgb (200GB) | Object Detection 遥感目标检测 |
Box | ObjEarth-Data/fMoW-rgb |
here | Optical satellite | ✅ | * |
| 2-1 | DIOR | Object Detection 遥感目标检测 |
Box | ObjEarth-Data/DIOR |
here | Optical satellite | ✅ | * |
| 2-2 | VEDAI | Vehicle Detection in Aerial Imagery 车辆检测 |
Box | ObjEarth-Data/VEDAI |
here | Optical satellite+Infrared | ✅ | ⬇️☑️ |
| 2-3 | OPMOD | Open Pit Mine Object Detection 露天矿区目标检测 |
Box | ObjEarth-Data/OPMOD |
here | Optical satellite | ✅ | ⬇️☑️ |
| 2-3 | MARIDA | Marine Debris Segmentation 海洋垃圾检测分割 |
Mask | ObjEarth-Data/MARIDA |
here | Multispectral (Sentinel-2) | ✅ | ⬇️☑️ |
| 2-3 | MADOS | Marine Debris and Oil Spill Segmentation 海洋垃圾和油污泄漏分割 |
Mask | ObjEarth-Data/MADOS |
here | Multispectral (Sentinel-2) | ✅ | ⬇️☑️ |
| 2-3 | FASDD_RS | Flame and Smoke Detection 火焰和烟雾检测 |
Box | ObjEarth-Data/FASDD_RS |
here | Optical satellite | ✅ | ⬇️☑️ |
| 1-* | DVGBench | UAV Visual Grounding 无人机视觉定位 |
Box+Referring | ObjEarth-Data/DVG_Bench/images/ - ObjEarth-Bench/DVGBench/ |
here | RGB | ✅ | ⬇️☑️ |
| 1-5 | AU-AIR | UAV Object Tracking and Detection (Traffic Surveillance) 交通监控下的目标跟踪检测 |
Box | ObjEarth-Data/AU-AIR/images/ |
RGB+Video | ⏳ | ⬇️☑️ | |
| 1-4 | uav_dataset | Person Detection Covered with Heavy Snow (Search and Rescue) 大雪覆盖下的人员探测搜救 |
Box | `` | here | RGB | ⏳ | ⬇️☑️ |
| 2-1 | RarePlanes (107GB) | Aircraft Detection 飞机检测 |
Box | `` | here | Optical satellite | * | * |
| 1-6 | MOBDrone (244GB) | Man OverBoard Rescue Detection 落水人员检测与救援 |
Box | `` | here | RGB+Video | * | * |
| 1-4 | DroneCrowd | Crowd Counting 人群目标定位/计数/跟踪/密度估计 |
Point | `` | here | Video | ⏳ | ⬇️☑️ |
| 1-4 | DroneRGBT | RGB-T Crowd Counting RGB-热成像人群计数 |
Point | `` | here | RGB-thermal | ⏳ | ⬇️☑️ |
| 1-4 | Up-Count | People Localisation and Counting 人群定位与计数 |
Point | `` | here | Video | * | ⬇️☑️ |
| 1-1 | MTC-UAV | Maize Tassels Counting 玉米穗粒计数 |
Point | `` | here | RGB | ⏳ | ⬇️☑️ |
| 1-* | RefDrone | Drone Object Referring 无人机目标指代理解 |
Box+Referring | ObjEarth-Data/VisDrone2019_DET/ - ObjEarth-Bench/RefDrone/ |
here | RGB | ✅ | ⬇️☑️ |
| 1-* | VisDrone2019_DET | Drone Object Detection 无人机目标检测 |
Box | ObjEarth-Data/VisDrone2019_DET/ |
RGB | ✅ | ⬇️☑️ | |
| 1-* | HazyDet | Drone Object Detection in Hazy 雾天无人机目标检测 |
Box | ObjEarth-Data/HazyDet/ |
RGB | ✅ | ⬇️☑️ | |
| 1-* | UDD | Aerial Semantic Segmentation 航拍语义分割 |
Mask | `` | here | RGB | * | ⬇️☑️ |
| 1-* | VDD | Aerial Semantic Segmentation 航拍语义分割 |
Mask | `` | here | RGB | * | ⬇️☑️ |
| 1-* | AeroScapes | Aerial Semantic Segmentatio 航拍语义分割 |
Mask | `` | here | RGB | * | ⬇️☑️ |
| 1-* | semantic_drone_dataset | Aerial Semantic Segmentation 航拍语义分割 |
Mask | `` | here | RGB | * | ⬇️☑️ |
| 1-4 | FloodNet | Aerial Semantic Segmentation & VQA 洪水语义分割/问答 |
Mask+VQA | `` | here | RGB | * | ⬇️☑️ |
| 1-* | UAVid | Aerial Semantic Segmentation 航拍语义分割 |
Mask | `` | here | RGB | * | ⬇️☑️ |
| 1-* | UAVDT-M | Single Object Tracking 单目标跟踪 |
Box | `` | here | Video | * | ⬇️☑️ |
| 1-* | UAVDT-S | Multiple Object Tracking 多目标跟踪 |
Box | `` | here | Video | * | ⬇️☑️ |
| 1-5 | DroneVehicle | RGB-Infrared Vehicle Detection RGB-红外车辆目标检测 |
Box | `` | here | RGB+Infrared | * | ⬇️☑️ |
| 1-* | UAV123 | Object Tracking 目标跟踪 |
Box | `` | Data source failed | Video | * | ⬇️☑️ |
| 1-* | DTB70 | Object Tracking 目标跟踪 |
Box | `` | here | Video | * | ⬇️☑️ |
| 1-5 | TAIVD | Traffic Aerial Images for Vehicle Detection 交通道路车辆检测 |
Box | `` | here or here | RGB | * | ⬇️☑️ |
| 1-* | WebUAV-3M (819G) | Object Tracking 无人机单目标跟踪 |
Box | `` | here | Video | * | ⬇️ |
| 1-* | MODA (100G) | Multispectral Object Detection 多光谱目标检测 |
Box | `` | here | Multispectral | * | ⬇️☑️ |
| 1-* | CODrone | Oriented Object Detection 无人机旋转目标检测 |
Box (Oriented) | `` | here | RGB | * | ⬇️☑️ |
| 1-* | UAVOD-10 | Small Weak Object Detection 小弱目标检测 |
Box | `` | here | RGB | * | ⬇️☑️ |
| 1-* | HIT-UAV | Thermal Infrared Object Detection 热红外目标检测 |
Box | `` | here here | Thermal | * | ⬇️☑️ |
| 1-* | RGBT-Tiny | Visible-Thermal Tiny Object Detection RGB–热红外小目标检测与跟踪 |
Box | `` | here | RGB-thermal+Video | * | ⬇️☑️ |
| 1-4 | Unicamp-UAV | Human Detection in UAV images 人体目标检测 |
Box | `` | here | RGB | * | ⬇️☑️ |
| 1-1 | BIRDSAI | Animals Detection and Tracking 热红外野生动物检测与跟踪 |
Box | `` | here | Thermal+Video | * | ⬇️☑️ |
| 1-1 | WAID | Wildlife Detection with Drones 野生动物检测 |
Box | `` | here | RGB | * | ⬇️☑️ |
| 1-5 | UAV-ROD | Car Detection 车辆检测 |
Box (Oriented) | `` | here | RGB | * | ⬇️☑️ |
| 1-* | AIRST-CS | RGB–infrared Object Detection RGB-红外目标检测 |
Box | `` | here | RGB+Infrared | * | ⬇️☑️ |
| 1-3 | MBDD2025 | Building Surface Defect Detection 建筑物表面缺陷检测 |
Box | `` | here here | RGB | * | ⬇️☑️ |
| 1-2 | DroneRail | Detection · segmentation · defect analysis · 3D perception 铁路综合巡检(检测/分割/缺陷分析/三维感知) |
Box+Mask+3D | `` | here (Do not share) | RGB | * | ❌ |
| 1-1 | dead_fish_dataset | Dead Fish UAV Detection in Aquaculture 水产养殖死鱼检测 |
Box | `` | here | RGB | * | ⬇️☑️ |
| 1-3 | BFD-UAV2K | UAV Facade Defect Detection 建筑外墙缺陷检测 |
Box | `` | here | RGB | * | ⬇️☑️ |
| 1-2 | DTU | Wind Turbine Blade Drone Inspection 风机叶片检测 |
Box | `` | here | RGB | * | ⬇️☑️ |
| 1-2 | Blade30 | blade stitching, segmentation, defect detection, classification and deduplication, contamination detection and classification 风力涡轮机叶片拼接/分割/缺陷检测/分类/污染检测分类 |
Box+Mask | `` | here | RGB | * | ⬇️☑️ |
| 1-2 | TPPD | Thermal PV Panel Detection 热成像光伏板检测与故障识别 |
Box | `` | here | Thermal | * | ⬇️☑️ |
| 1-2 | UAV-RSOD | Railroad Obstacle Detection 铁路障碍物分割与检测 |
Box+Mask | `` | here | RGB | * | ⬇️☑️ |
| 1-1 | CampanetaOrange | Orange Fruit Detection in Citrus Orchards 柑桔果实检测 |
Box | `` | here | Multispectral | * | ⬇️ |
| 1-1 | PWD-UAV | identifying diseased pine trees affected by pine wood nematode 受松木线虫影响的病害松树识别 |
Box | `` | here | Multispectral | * | * |
| 1-1 | ICAERUS_Cow | Cow Detection and Counting 牛群检测跟踪计数 |
Box | `` | here | RGB+Video | * | * |
| 1-1 | ICAERUS_Sheep | Sheep Detection and Counting 羊群检测跟踪计数 |
Box | `` | here | RGB+Video | * | * |
| 1-1 | ICAERUS_Goat | Goats/Small Ruminants Detection and Counting 山羊及小反刍动物检测跟踪计数 |
Box | `` | here | RGB+Video | * | * |
| 1-2 | De-Solar | Obstruction Localization in solar PV systems 太阳能光伏系统中遮挡物定位 |
Mask | `` | here | RGB | * | * |
| 1-1 | ADD256 | Apples by Drones Detection 苹果检测 |
Box (from Circle) | `` | here | RGB | * | ⬇️☑️ |
| 1-2 | UAV-CrackX | Pavement Crack Semantic Segmentation 路面裂缝语义分割 |
Mask | `` | here | RGB | * | * |
| 1-5 | DroneWaste | Waste Recognition in Drone Imagery 垃圾识别 |
Box+Mask | `` | here | RGB | * | * |
| 1-3 | CUBIT-InSeg | Façade Defect Segmentation 建筑外立面缺陷检测 |
Mask | `` | here | RGB | * | * |
| 1-3 | CUBIT-Det | Infrastructure Defect Detection 基础设施(建筑、桥梁、路面)缺陷检测 |
Box | `` | here | RGB | * | ⬇️ |
| 1-1 | MOPAD | Growing Status Observation for Oil Palm Tree 油棕树长势观测 |
Box | `` | here | RGB | * | * |
| 1-1 | OPTD | Oil Palm Tree Detection for Anomaly Identification 油棕树异常检测 |
Box | `` | here | RGB | * | ⬇️☑️ |
| 1-* | MAVREC | Multiview Aerial Visual Recognition 地面-无人机多视角目标检测 |
Box | `` | here | Video | * | * |
| 1-* | VTUAV | Visible-Thermal UAV Tracking 可见光—热红外跟踪 |
Box | `` | here | RGB-thermal | * | * |
| 1-4 | UAV_piping | Piping Detection of River Embankments 无人机河岸堤防管涌检测 |
Box (within image) | `` | here | RGB+Infrared | * | ⬇️☑️❌ |
| 1-* | SDD (Stanford Drone Dataset) | Target Tracking or Trajectory Forecasting 校园人员跟踪与轨迹预测 |
Box | `` | here | Video | * | ❌ |
| 1-* | UAV-Human | Human Behavior Understanding 人体行为/姿态/动作理解 |
Action Class | `` | here | RGB+Infrared (Depth-Skeleton- Fisheye-Night-vision) |
* | ❌ |
| 1-* | UAVDark135 | UAV Nighttime Tracking 夜间单目标跟踪 |
Box | `` | here | RGB | * | ❌ |
| 1-4 | FireMan-UAV-RGBT | Wildfire Detection 针叶林火灾检测 |
Box | `` | here | RGB-thermal+Video | * | * |
| 1-* | AeroEye (ERA、MAVREC) | Multi-Object Relationship Modeling 视频目标关系建模 |
Box | `` | here | Video | * | * |
| 1-* | MUST | Multispectral UAV Single Object Tracking 多光谱单目标跟踪 |
Box | `` | here | Multispectral | * | * |
| 1-4 | Manipal-UAV Person | Person Detection 人类目标检测 |
Box | `` | here | RGB | * | * |
| 1-* | msuav500k | Foundational Dataset for Multispectral and RGB UAV Imagery 大规模无人机多光谱和RGB图像 |
None | `` | here | Multispectral+RGB | * | * |
| 1-3 | MineCareerDB | Mining Career Analysis 露天矿山巡检 |
Box | `` | here | RGB | * | * |
| 1-6 | Coastal_Litter | Coastal Litter Detection 海岸带垃圾检测 |
None | `` | here | RGB+Thermal+Multispectral | * | * |
| 1-2 | AQUADA-GO | Blade Damage Detection and Risk Evaluation 风机叶片损伤检测 |
None | `` | here | RGB-thermal+Video | * | * |
| 1-* | WildUAV | Depth Estimation and Semantic Segmentation 单目深度估计和语义分割 |
Mask+depth | `` | here | RGB | * | * |
| 1-* | MONET | Drone Thermal Dataset in Rural Scenarios 农村热红外目标检测 |
Box | `` | here | Thermal | * | * |
| 1-1 | BaleUAVision | Hay Bale Detection and Counting 农田干草捆检测与计数 |
Box+Mask | `` | here | RGB | * | * |
| 1-1 | PDT | Target Detection for Pests and Diseases Tree 病虫害树木检测 |
Box | `` | here | RGB | * | * |
| 1-1 | GWHD_2021 | Wheat Head Detection 小麦穗检测 |
Box | `` | here here | RGB | * | * |
| 1-1 | MMLA_OPC | Animal (zebras&giraffe) Detection 斑马\长颈鹿监测 |
Box | `` | here | Video | * | * |
| 1-1 | MMLA-Wilds | Wildlife Detection 斑马\长颈鹿\犬监测 |
Box | `` | here | Video | * | * |
| 1-1 | WildLive | Wildlife Tracking 野生动物跟踪(斑马\长颈鹿\大象) |
Box+Mask+Trajectory | `` | here | Video | * | * |
| 1-1 | BaboonLand | Tracking Primates in the Wild 狒狒行为与群体活动监测 |
Box+Behavior | `` | here | Video | * | * |
| 1-2 | VEPL | Segmentation of Vegetation Encroachment in Power Line Corridors 输电线路走廊植被侵入分割 |
Mask | `` | here | RGB | * | * |
| 1-2 | PLD-UAV | Power Line Detection 电力线路分割 |
Mask | `` | here | RGB | * | * |
| 1-1 | UAV-WTD | Waterfowl Thermal Detection 无人机热红外水鸟监测 |
Box | `` | here | Thermal | * | * |
| 1-1 | WeedMap | Weed/Crop Segmentation 多光谱作物\杂草分割 |
Mask | `` | here | Multispectral | * | * |
| 1-1 | TTADDA-UAV | Multi-Season RGB and Multispectral Dataset of Potato Fields 多季节RGB\多光谱马铃薯监测 |
None | `` | here here | RGB+Multispectral | * | * |
| 2-1 | xView3-SAR (~3TB) | Ship Detection(dark vessels) 船舶检测 |
Box | `` | here | SAR | ❌ | ❌ |
| 2-1 | fMoW-full (3.5TB) | Object Detection 遥感目标检测 |
Box | `` | here | Multispectral | ❌ | ❌ |
| 2-1 | SatlasPretrain (~33TB) | large-scale pre-training dataset for remote sensing image understanding 大规模遥感图像理解 |
Point+Mask | `` | here | Optical satellite | ❌ | ❌ |
| 1-* | UAV | Object | Box | | [here]() | RGB | * | * | | 1-* | UAV | Object | Box | | here | RGB | * | * |
| 序号 | 无人机 Domain | 领域定位 | 典型检测目标 | 典型应用 | 核心输出 / 特点 |
|---|---|---|---|---|---|
| 1-1 | UAV-农林牧业与生态监测 | 检测自然环境中的生物目标及农业生产对象 | 农作物、果实、花朵、植株 缺苗、倒伏作物、杂草 牛、羊等牲畜 野生动物、鸟类 病树、枯死树、倒伏树木 农业机械、作业人员 |
作物株数和果实数量统计 缺苗、杂草和倒伏区域发现 牲畜自动盘点和走失牲畜搜索 野生动物检测、跟踪和数量调查 森林病树、枯死树和风倒木识别 农林区域非法人员或车辆检测 |
目标类别 目标数量 目标位置 目标密度 变化趋势 |
| 1-2 | UAV-能源、工业与基础设施巡检 | 将电力、油气、工厂、桥梁、道路等统一归入资产和设施巡检 | 输电杆塔、导线、绝缘子、金具 鸟巢、异物、悬挂物 光伏组件、热斑组件 风机叶片、裂纹及表面损伤 管道、阀门、储罐、烟囱 桥梁构件、道路设施、铁路设备 裂缝、锈蚀、破损、脱落 设备热点、泄漏点和异常部件 |
电力线路异物和缺陷检测 绝缘子、连接件和杆塔部件检测 光伏板热斑及故障组件定位 风机叶片损伤检测 油气管线和工业设备异常检测 桥梁、道路、铁路设施缺陷识别 设备缺失、错位或损坏检测 |
部件检测:定位设备及零部件 缺陷检测:识别裂纹、腐蚀、破损、过热等异常 |
| 1-3 | UAV-建筑、施工与矿山作业 | 将施工现场、露天矿、采石场和大型工程现场视为复杂作业区域 | 施工人员 安全帽、反光衣、安全绳等防护装备 挖掘机、吊车、卡车、推土机 建筑材料和堆料 临时建筑、围栏和危险区域 裂缝、坍塌、落石 非法进入人员和车辆 烟雾、明火和积水 |
施工人员和作业车辆统计 安全帽、反光衣等PPE检测 人员进入危险区域检测 起重设备周边人员检测 矿区车辆及机械设备调度监测 边坡落石、裂缝和坍塌检测 非法采矿、非法占地目标发现 施工现场火灾和烟雾检测 |
判断现场有哪些目标 确定目标所在位置 判断人员、车辆或设备是否违反安全规则 |
| 1-4 | UAV-公共安全、应急与安防 | 将消防、搜救、灾害响应、边境和园区安防整合为安全应用领域 | 人员、失踪者、受困人员 车辆、船只 明火、烟雾和高温热点 聚集人群 越界人员 可疑车辆或可疑物体 灾区生命目标 夜间移动目标 |
山区、森林和水域搜救 红外人体检测 火点、烟雾和复燃点检测 洪水、地震等灾害后的人员和车辆搜索 园区、边境及重要设施周界巡逻 人群聚集和异常活动检测 夜间入侵人员检测 危险区域非法进入检测 |
红外相机可用于夜间人员检测 烟雾遮挡环境中的热目标发现 森林或复杂地形中的生命目标搜索 火灾热点定位 |
| 1-5 | UAV-交通与智慧城市 | 关注道路、城市空间及交通系统中的动态目标检测 | 小汽车、卡车、公交车 摩托车、自行车 行人 停放车辆 交通事故车辆 道路障碍物 抛洒物、落石 占道施工设备 |
车辆检测、分类和计数 车流量和道路拥堵分析 车辆轨迹和速度分析 违停、逆行和异常停车检测 交通事故自动发现 道路障碍物和抛洒物检测 大型活动人群数量统计 停车场车位占用检测 |
目标检测负责发现目标 多目标跟踪负责持续识别同一目标 可进一步计算车速、轨迹、停留时间和交通流量 |
| 1-6 | UAV-海洋水域与环境执法 | 将海岸、河流、湖泊、港口及海上监管统一归入水域监测 | 船舶、渔船和小型快艇 漂浮人员和救生设备 海洋动物 漂浮垃圾 水面油污和异常漂浮物 非法采砂船 非法捕捞船只 河道及岸边非法排放设施 |
船舶检测、分类和数量统计 非法捕捞及越界船只检测 水上搜救和落水人员检测 海洋动物检测和数量调查 河流、湖泊漂浮垃圾检测 水面异常目标和污染物发现 港口和海岸线安防巡逻 非法采砂、非法排放行为发现 |
重点解决水面反光、波浪干扰和远距离小目标问题 可结合长焦、红外和视频跟踪提升检测效果 |
| 序号 | 卫星 Domain | 典型检测目标 | 多光谱数据与公开数据集依据 | 典型任务与输出 |
|---|---|---|---|---|
| 2-1 | Sat-海事、航空与交通目标监测 | 海事目标:船舶、渔船、大型货船、海上平台及港口目标 航空目标:飞机、直升机及机场地面航空器 飞机细粒度类型:Boeing 737、747、777、787,Airbus A220、A321、A330、A350,ARJ21、C919及其他飞机类型 飞机属性:机身长度、翼展、机翼形状、机翼位置、动力类型、发动机数量、垂尾数量及飞机用途 交通设施:机场、跑道、停机坪、港口、船厂、道路桥梁和铁路桥梁 |
S2-SHIPS使用Sentinel-2多光谱影像开展船舶检测,说明船舶是原生多光谱目标检测中相对成熟的方向。(MDPI) fMoW-full提供4波段和8波段可见光—近红外影像,可支持机场、港口、船厂和交通设施等目标级识别。(GitHub) FAIR1M提供旋转框细粒度遥感目标检测,明确要求模型不仅检测“飞机”,还要识别Airbus A350、Boeing 747等具体类型。(ar5iv) RarePlanes提供飞机检测以及10类细粒度结构属性标注。(arXiv) |
船舶和飞机定位、分类、计数 飞机具体型号识别 飞机结构属性与用途分类 飞机停放位置及朝向检测 船舶尺寸、方向和航行状态估计 机场、港口及交通枢纽活动监测 |
| 2-2 | Sat-城市建筑与公共基础设施 | 建筑物、住宅区和大型厂房 施工场地及临时建筑 机场和公共服务设施 道路、桥梁、水坝和交通节点 受损、倒塌或被洪水影响的建筑物 |
fMoW-full的多光谱数据覆盖住宅、施工场地、医院、学校、公共机构、桥梁和水坝等设施类别。(GitHub) SatlasPretrain融合Sentinel-2与高分辨率航片,提供建筑、机场、道路等多种点、线和多边形标注。(CVF Open Access) |
建筑物和基础设施定位 设施数量及空间分布统计 新增建筑和施工活动发现 公共设施识别 灾前灾后建筑目标对比 |
| 2-3 | Sat-能源、工业与资源设施 | 光伏电站和光伏阵列 风力发电机及风电场 储油罐、储罐区和炼化设施 发电厂、变电站及工业园区 海上油气平台 矿区、采石场和资源开发设施 水处理和垃圾处理设施 |
fMoW-full包含光伏电站、风电场、发电厂、变电站、油气设施、储罐、露天矿和水处理设施等多光谱目标类别。(GitHub) SatlasPretrain包含风机点目标、光伏场区域目标、建筑和其他基础设施标注,并使用Sentinel-2等卫星数据。(CVF Open Access) |
能源资产检测和计数 工业设施定位与分类 风机、光伏场和海上平台普查 储罐及大型设备统计 资源开发活动和设施变化监测 |