FLASH：无人机滑坡灾情轻巧感知系统

doi:10.7527/S1000-6893.2025.31293

Abstract

Abstract:

To address the urgent needs of emergency response for major landslide disasters under extreme conditions such as disconnected roads， power outages， and network failures， this study designs and develops FLASH， a lightweight UAV-based landslide perception system which integrates hardware and algorithms. The hardware system includes a lightweight perception module and real-time interpretation unit， designed as decoupled standalone nodes. Utilizing ROS-based messaging and serial communication for low-latency data transfer， the system operates at a peak power of 25 W and weighs 500 g. Algorithmically， a lightweight design model incorporating spatial-informational constraints is proposed， which maintains inference accuracy while drastically reducing latency for semantic segmentation of ultra-high-resolution images （e.g.， 78.88 ms per 7 952×5 304 image）. The model further enables intelligent multi-image landslide retrieval and volume estimation via deep semantic representation， achieving 81% reliability in volume calculations （validated in the Songpan landslide case） without relying on 3D reconstruction. Ultimately， this system realizes the fully automated extraction of critical disaster information from UAV imagery acquisition to landslide volume estimation， significantly enhancing the efficiency of landslide disaster interpretation.

Key words: real-time landslide extraction, UAV imagery, emergency response, edge device, system integration, lightweight design, ROS system

CLC Number:

V279

Lin CHEN, Qing ZHU, Han HU, Yulin DING, Pengxin GU. FLASH: Flexible and lightweight awareness of slope hazard[J]. Acta Aeronautica et Astronautica Sinica, 2025, 46(11): 531293.

Figures/Tables 17

Fig.1

Fig.2

Fig.3

Fig.4

Fig.5

Fig.6

Fig.7

Fig.8

Fig.9

Table 1

Table 2

Table 3

Fig.10

Table 4

Table 5

Fig.11

Fig.12

References

	‍［1］朱庆，曾浩炜，丁雨淋，等. 重大滑坡隐患分析方法综述‍［J］. 测绘学报， 2019， 48（12）： 1551-1561.
	ZHU Q， ZENG H W， DING Y L， et al. A review of major potential landslide hazards analysis‍‍［J］. Acta Geodaetica et Cartographica Sinica， 2019， 48（12）： 1551-1561 （in Chinese）.
	‍［2］ NIYOKWIRINGIRWA P， LOMBARDO L， DEWITTE O， et al. Event-based rainfall-induced landslide inventories and rainfall thresholds for Malawi‍［J］. Landslides， 2024， 21（6）： 1403-1424.
	‍［3］ WANG X， FAN X M， XU Q， et al. Change detection-based co-seismic landslide mapping through extended morphological profiles and ensemble strategy‍［J］. ISPRS Journal of Photogrammetry and Remote Sensing， 2022， 187： 225-239.
	‍［4］ ZHOU M J， YUAN M Z， YANG G X， et al. Risk analysis of road networks under the influence of landslides by considering landslide susceptibility and road vulnerability： A case study‍［J］. Natural Hazards Research， 2024， 4（3）： 387-400.
	‍［5］ CHEN L， DING Y L， HU H， et al. Landslide extraction using fused local and nonlocal attentional features on edge device toward embedded UAV emergency response‍［J］. IEEE Transactions on Geoscience and Remote Sensing， 2024， 62： 5625720.
	‍［6］朱庆，曹振宇，林珲，等. 应急测绘保障体系若干关键问题研究‍［J］. 武汉大学学报（信息科学版）， 2014， 39（5）： 551-555.
	ZHU Q， CAO Z Y， LIN H， et al. Key technologies of emergency surveying and mapping service system‍［J］. Geomatics and Information Science of Wuhan University， 2014， 39（5）： 551-555 （in Chinese）.
	‍［7］刘飞，朱庆，丁雨淋，等. 滑坡—堰塞湖灾情无人机应急测绘、分析与险情模拟‍［J］. 山地学报， 2021， 39（4）： 600-610.
	LIU F， ZHU Q， DING Y L， et al. Analysis and simulation of landslide-barrier lake disaster based on UAV emergency mapping‍［J］. Mountain Research， 2021， 39（4）： 600-610 （in Chinese）.
	‍［8］ WAN Y T， ZHONG Y F， MA A L， et al. An accurate UAV 3-D path planning method for disaster emergency response based on an improved multiobjective swarm intelligence algorithm‍［J］. IEEE Transactions on Cybernetics， 2023， 53（4）： 2658-2671.
	‍［9］ ZHANG Y J， LI J， LIU J J， et al. Extraction of landslide morphology based on topographic profile along the direction of slope movement using UAV images‍［J］. Geomatics， Natural Hazards and Risk， 2023， 14（1）： 2278276.
	‍［10］ LIU P， WEI Y M， WANG Q J， et al. Research on post-earthquake landslide extraction algorithm based on improved U-Net model‍［J］. Remote Sensing， 2020， 12（5）： 894.
	‍［11］ YAO Z H， CHENG W C， ZHANG W， et al. The rise of UAV fleet technologies for emergency wireless communications in harsh environments‍［J］. IEEE Network， 2022， 36（4）： 28-37.
	‍［12］ LONG J， SHELHAMER E， DARRELL T. Fully convolutional networks for semantic segmentation‍［C］∥2015 IEEE Conference on Computer Vision and Pattern Recognition（CVPR）. Piscataway： IEEE Press， 2015： 3431-3440.
	‍［13］ BO W H， LIU J， FAN X J， et al. BASNet： Burned area segmentation network for real-time detection of damage maps in remote sensing images‍‍［J］. IEEE Transactions on Geoscience and Remote Sensing， 2022， 60： 5627913.
	‍［14］ LI M L， ZHAO X K， LI J S， et al. ComNet： Combinational neural network for object detection in UAV-borne thermal images‍［J］. IEEE Transactions on Geoscience and Remote Sensing， 2020， 59（8）： 6662-6673.
	‍［15］ SANDRIC I， CHITU Z， ILINCA V， et al. Using high-resolution UAV imagery and artificial intelligence to detect and map landslide cracks automatically‍［J］. Landslides， 2024， 21（10）： 2535-2543.
	‍［16］ DONG Z Y， AN S， ZHANG J， et al. L-unet： A landslide extraction model using multi-scale feature fusion and attention mechanism‍［J］. Remote Sensing， 2022， 14（11）： 2552.
	‍［17］ CHEN H S， HE Y， ZHANG L F， et al. A landslide extraction method of channel attention mechanism U-Net network based on Sentinel-2A remote sensing images‍［J］. International Journal of Digital Earth， 2023， 16（1）： 552-577.
	‍［18］ JI S P， YU D W， SHEN C Y， et al. Landslide detection from an open satellite imagery and digital elevation model dataset using attention boosted convolutional neural networks‍［J］. Landslides， 2020， 17（6）： 1337-1352.
	‍［19］ MACENSKI S， FOOTE T， GERKEY B， et al. Robot Operating System 2： Design， architecture， and uses in the wild‍［J］. Science Robotics， 2022， 7（66）： eabm6074.
	‍［20］ CHEN Y P， DAI X Y， CHEN D D， et al. Mobile-former： Bridging MobileNet and transformer‍［C］∥2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition （CVPR）. Piscataway： IEEE Press， 2022.
	‍［21］ MEHTA S， RASTEGARI M. MobileViT： Light-weight， general-purpose， and mobile-friendly vision transformer‍［C］∥International Conference on Learning Representations， 2022.
	‍［22］ DEHGHANI M， ARNAB A， BEYER L， et al. The efficiency misnomer‍［DB/OL］. arXiv preprint： 2110.12894， 2022.
	‍［23］ VASU P K A， GABRIEL J， ZHU J， et al. MobileOne： an improved one millisecond mobile backbone‍［C］∥2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition （CVPR）. Piscataway： IEEE Press， 2023.
	‍［24］ DING X H， ZHANG X Y， MA N N， et al. RepVGG： Making VGG-style ConvNets great again‍［C］‍∥2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition （CVPR）. Piscataway： IEEE Press， 2021.
	‍［25］ Vaswani A. Attention is all you need‍［C］∥Advances in Neural Information Processing Systems， 2017.
	‍［26］ FU J， LIU J， TIAN H J， et al. Dual attention network for scene segmentation‍［C］‍∥2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition （CVPR）. Piscataway： IEEE Press， 2019.
	‍［27］ KATHAROPOULOS A， VYAS A， PAPPAS N， et al. Transformers are RNNs： Fast autoregressive transformers with linear attention‍［C］∥International conference on machine learning， 2020.
	‍［28］ HAN D C， YE T Z， HAN Y Z， et al. Agent attention： On the integration of softmax and linear attention‍［M］∥Computer Vision-ECCV 2024. Cham： Springer Nature Switzerland， 2024.
	‍［29］ XIE E Z， WANG W H， YU Z D， et al. SegFormer： Simple and efficient design for semantic segmentation with transformers‍［J］. Advances in Neural Information Processing Systems， 2021， 34： 12077-12090.
	‍［30］ GUO M H， LU C Z， HOU Q， et al. Segnext： Rethinking convolutional attention design for semantic segmentation‍［J］. Advances in Neural Information Processing Systems， 2022， 35： 1140-1156.
	‍［31］方群生，唐川，程霄，等. 汶川震区泥石流流域内滑坡物源量计算方法探讨‍［J］. 水利学报， 2015， 46（11）： 1298-1304.
	FANG Q S， TANG C， CHENG X， et al. An calculation method for predicting landslides volumes of the debris flows in the Wenchuan earthquake area‍［J］. Journal of Hydraulic Engineering， 2015， 46（11）： 1298-1304 （in Chinese）.
	‍［32］ RONNEBERGER O， FISCHER P， BROX T. U-Net： Convolutional networks for biomedical image segmentation‍［M］‍∥Medical Image Computing and Computer-Assisted Intervention-MICCAI 2015. Cham： Springer International Publishing， 2015： 234-241.
	‍［33］ CHEN L C， ZHU Y K， PAPANDREOU G， et al. Encoder-decoder with atrous separable convolution for semantic image segmentation‍［C］∥Computer Vision-ECCV 2018. Cham： Springer International Publishing， 2018.
	‍［34］ YAN S， WU C L， WANG L Z， et al. DDRNet： Depth map denoising and refinement for consumer depth cameras using cascaded CNNs‍［C］∥Computer Vision-ECCV 2018. Cham： Springer International Publishing， 2018.
	‍［35］ YU C Q， WANG J B， PENG C， et al. BiSeNet： Bilateral segmentation network for real-time semantic segmentation‍［C］‍∥Computer Vision-ECCV 2018. Cham： Springer International Publishing， 2018： 334-349.
	‍［36］ LI X， ZHOU Y M， PAN Z， et al. Partial order pruning： For best speed/accuracy trade-off in neural architecture search‍［C］∥2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition （CVPR）. Piscataway： IEEE Press， 2019.
	‍［37］ LI X T， YOU A S， ZHU Z， et al. Semantic flow for fast and accurate scene parsing‍［M］∥Computer Vision-ECCV 2020. Cham： Springer International Publishing， 2020： 775-793.

序号	地区	已标注影像数量	影像尺寸/像素	采集时间	传感器类型
1	茂县新磨村	49	6 000×4 000	2017-06-25	SONY ILCE-5100
2	九寨沟树正村	74	6 000×4 000	2017-08-13	SONY ILCE-5100
3	九寨沟荷叶寨	53	6 000×4 000	2017-08-14	SONY ILCE-5100
4	九寨沟上四寨村	50	6 000×4 000	2017-08-16	SONY ILCE-5100
5	汶川阿尔寨	75	5 472×3 468	2018-04-18	DJI FC6310
6	贵州六盘水发耳	82	5 472×3 468	2018-07-26	DJI FC6310S
7	金沙江白格	83	8 688×5 792	2018-11-05	Canon EOS 5DS
8	泸定磨西镇	55	7 952×5 304	2022-09-07	SONY DSC-RX1RM2
9	阿坝州松潘县	55	7 952×5 304	2023-05-12	SONY ILCE-7RM2

方案（作业流程）	工作模式	依赖设备	解译数据类型	影像尺寸/像素	数据采集时间	三维重建时间			解译时间
方案（作业流程）	工作模式	依赖设备	解译数据类型	影像尺寸/像素	数据采集时间	影像对齐	密集匹配	生成DOM	解译时间
基于GF2卫星影像DOM	离线	图形工作站	卫星影像（DOM）	2 542×2 941	36 d 19 h 47 min 12 s				20 s
基于无人机影像DOM	离线	高性能图形工作站	无人机影像（DOM）	25 237×48 421	20 min 18 s	3 h 45 min 4 s	1 d 5 h 40 min 2 s	2 min 1 s	33 min 41 s
本文方案	实时	边缘设备	无人机影像	7 952×5 304					78.88 ms （×366）

方法	影像尺寸	参数量/M	FLOPs /G	推理延迟/ms		IoU/%
方法	影像尺寸	参数量/M	FLOPs /G	GPU	边缘设备	IoU/%
MobileOne-s1^［23］	1 024×1 024	5.13	17.25	15.92	355.55	46.17
DDRNet-23-Slim^［34］	1 024×1 024	5.70	18.32	11.88	306.79	59.50
RepVGG-A0^［24］	1 024×1 024	8.67	28.46	10.54	317.07	53.57
DF1-Seg^［36］	1 024×1 024	9.37	4.67	25.21	68.33	31.90
BiSeNetV1^［35］	1 024×1 024	13.42	61.04	20.93	624.52	59.24
SFNet（ResNet18）^［37］	1 024×1 024	12.87	121.89	42.23		54.27
RepVGG-A1^［24］	1 024×1 024	13.16	49.41	15.83	371.05	56.00
RepVGG-B0^［24］	1 024×1 024	14.70	63.90	20.08	521.60	50.43
FCN^［12］	1 024×1 024	18.64	407.98	218.84	2 439.48	38.09
DF2-Seg^［36］	1 024×1 024	18.82	12.18	12.70	136.19	33.77
DDRNet-23^［34］	1 024×1 024	20.15	71.88	26.11	640.91	41.59
RepVGG-A2^［24］	1 024×1 024	26.08	106.94	29.52	772.28	55.02
UNet^［32］	1 024×1 024	34.53	1 048.36	436.59	2 095.92	31.39
DeeplabV3+^［33］	1 024×1 024	59.23	354.14	215.09	2 485.54	59.87
DDRNet-39^［34］	1 024×1 024	32.36	141.08	45.50	1 139.68	60.46
RepVGG-B1^［24］	1 024×1 024	53.99	246.95	63.31	1 440.39	59.54
RepVGG-B2^［24］	1 024×1 024	84.32	384.08	94.35	2 743.62	57.19
RepVGG-B3^［24］	1 024×1 024	114.97	547.76	131.49	3 283.37	53.74
RepFlash	1 024×1 024	5.39	76.79	37.95	78.88	67.73

类别	主干网络	局部特征	全局特征	参数量/M	FLOPs/G	IoU/%
A	MobileOne	×	×	1.06	5.75	58.07
B	MobileOne	×	DA Block	28.66	458.73	58.26
	MobileOne	×	LA Block	4.20	57.28	61.42
	MobileOne	×	LGM Block	4.21	57.46	62.26
C	MobileOne	√	×	2.24	25.07	59.59
	MobileOne	√	DA Block	28.92	463.03	64.56
	MobileOne	√	LA Block	5.38	76.61	62.23
	MobileOne（本文模块）	√	LGM Block	5.39	76.79	67.73

结构	参数量/M	FLOPs/G	GPU延迟/ms	边缘设备延迟/ms	IoU/%
RepFlash+a	5.26	74.64（-2.15）	36.56	80.15	62.06
RepFlash+b	5.71	310.36（+233.57）	152.14	403.32	65.96
RepFlash+c	5.66	81.08（+4.29）	40.19	84.52	60.41
RepFlash+d	5.39	76.79	37.95	78.88	67.73

FLASH: Flexible and lightweight awareness of slope hazard

RichHTML

PDF (PC)

Knowledge

Abstract

Cite this article

share this article

Figures/Tables 17

References

Related Articles 9

Recommended Articles

Metrics

Comments

[1]	Chuanyan FENG, Shuang LIU, Xiaoru WANYAN, Menglong DING, Daochun LI, Yaoming ZHOU. Research progress on key technologies for human factors design of unmanned aircraft systems [J]. Acta Aeronautica et Astronautica Sinica, 2025, 46(11): 531213-531213.
[2]	Guirong ZHOU, Jianyuan XU, Shaobo MA, Junyao ZONG, Jinqing SHEN, Haijie ZHU. Review of key technologies for avionics systems integration on large passenger aircraft [J]. Acta Aeronautica et Astronautica Sinica, 2024, 45(5): 529956-529956.
[3]	Weihong ZHANG, Changhong TANG. Lightweighting of aerospace and aeronautical equipment: Challenges and perspectives [J]. Acta Aeronautica et Astronautica Sinica, 2024, 45(5): 529965-529965.
[4]	Guodong XU, Danlei ZHANG, Zhendong XU. Arrival time processing method of pulsar characteristic frequency signals [J]. ACTA AERONAUTICAET ASTRONAUTICA SINICA, 2023, 44(3): 526185-526185.
[5]	JIANG Bo, QU Ruokun, LI Yandong, LI Chenglong. Object detection in UAV imagery based on deep learning: Review [J]. ACTA AERONAUTICAET ASTRONAUTICA SINICA, 2021, 42(4): 524519-524519.
[6]	SHEN Zhengyang, CHEN Xiaoming, HUANG Lingcai. Challenges for aircraft design due to special mission models of large-scale amphibious aircraft [J]. ACTA AERONAUTICAET ASTRONAUTICA SINICA, 2019, 40(1): 522400-522400.
[7]	ZHANG Zhenhui, LI Dong, YANG Yin. Passive flow control of multi-element airfoils using slat mini-trailing edge device [J]. ACTA AERONAUTICAET ASTRONAUTICA SINICA, 2017, 38(5): 120650-120650.
[8]	Zhang Weihong;Yang Jungang;Zhu Jihong. Simultaneous Topology and Shape Optimization of Pressure Loaded Structures [J]. ACTA AERONAUTICAET ASTRONAUTICA SINICA, 2009, 30(12): 2335-2341.
[9]	WANG Gang-lin;WU Zhe. System Frame Research on Aircraft Conceptual Design Platform Based on Virtual Prototyping [J]. ACTA AERONAUTICAET ASTRONAUTICA SINICA, 2005, 26(2): 162-167.