LARD 2.0: Enhanced Datasets and Benchmarking for Autonomous Landing Systems

Yassine Bougacha; Geoffrey Delhomme; Mélanie Ducoffe; Augustin Fuchs; Jean-Brice Ginestet; Jacques Girard; Sofiane Kraiem; Franck Mamalet

cs.ROMar 23, 2026

LARD 2.0: Enhanced Datasets and Benchmarking for Autonomous Landing Systems

Authors: Yassine Bougacha, Geoffrey Delhomme, Mélanie Ducoffe, Augustin Fuchs, Jean-Brice Ginestet, Jacques Girard, Sofiane Kraiem, Franck Mamalet,

Organizations: DGA · IRT Saint Exupéry, Toulouse, France · Thales, Toulouse, France · Airbus, Toulouse, France ·

Abstract

This paper addresses key challenges in the development of autonomous landing systems, focusing on dataset limitations for supervised training of Machine Learning (ML) models for object detection. Our main contributions include: (1) Enhancing dataset diversity, by advocating for the inclusion of new sources such as BingMap aerial images and Flight Simulator, to widen the generation scope of an existing dataset generator used to produce the dataset LARD; (2) Refining the Operational Design Domain (ODD), addressing issues like unrealistic landing scenarios and expanding coverage to multi-runway airports; (3) Benchmarking ML models for autonomous landing systems, introducing a framework for evaluating object detection subtask in a complex multi-instances setting, and providing associated open-source models as a baseline for AI models' performance.

Figures & tables

Along-track distance (m)	Yaw range (°)	Roll range (°)
$[-6000,\,-4500]$	$[-24,\;24]$	$[-30,\;30]$
$[-4500,\,-2500]$	$[-24,\;24]$	$[-15,\;15]$
$[-2500,\,-280]$	$[-18.5,\;18.5]$	$[-10,\;10]$

Table 1: Allowed yaw and roll ranges as a function of along-track distance from the LTP

Figure 1: Distribution of runways per airports in LARD v2

Figure 2: VBL constituent architecture with 3 stages

Figure 3: Lard V2 scenario generation and labeling workflow.

Input image RGB	$Img$	shape (H,W,3)
Runway BB	$\mathcal{B}_{i}=(x_{i},y_{i},w_{i},h_{i})$	$(x_{i},y_{i})$ center of the BB
Runway ODD indices (or flag)	$o_{i}\in\{0,1\}$	indicate if the runway is in the ODD
Predicted BB	$\hat{\mathcal{B}}_{i}=(\hat{x}_{i},\hat{y}_{i},\hat{w}_{i},\hat{h}_{i},\hat{s}_{i})$	$\hat{s}_{i}$ is confidence score
List of In ODD runways	$I=\{i\|o_{i}=1\}$
List of out-of-ODD runways	$J=\{j\|o_{j}=0\}$
List of predicted runways	$P=\{i\|\hat{s}_{i}>0.5\}$	can also use the NMS (Non Maximum Suppression) algorithm

Figure 5: Notations to define mAP and extended mAP

Model: archi (Trained on)	Score	Test GT	GT	Detections	mAP	mAP50	mAP75
$M_{IN}$ : yoloV8 (IN_ODD)	mAP	IN_ODD	53610	80002	0.6974	0.9671	0.8122
	mAP	IN_ODD + EXTENDED_ODD	61788	80002	0.6498	0.9084	0.7539
	e-mAP	IN_ODD + EXTENDED_ODD	58242	80002	0.6882	0.9654	0.7999
$M_{EXT}$ : yoloV8 (IN_ODD+ EXTENDED_ODD)	mAP	IN_ODD	53610	100504	0.6265	0.8703	0.7256
	mAP	IN_ODD + EXTENDED_ODD	61788	100504	0.6939	0.9669	0.8052
	e-mAP	IN_ODD + EXTENDED_ODD	61418	100504	0.6976	0.9685	0.8072

Table 2: Performance comparison between a yolov8 trained on In ODD runways only, and one trained on In ODD and Extended ODD runways.

Figure 6: Comparison of mAP@50 performances for several training setup

Panel of Figure 7 — Figure 7: Crossbar mAP performances of the eleven trained YOLOv11 models on each sub-test set.

Aerial Robotics

Aerial robotics studies the design, perception, control, and coordination of flying robots for navigation, inspection, transport, and physical interaction.

See all 340 papers

SafeLand: Safe Autonomous Landing in Unknown Environments with Bayesian Semantic Mapping
Mar 18, 2026Markus Gross, Andreas Greiner, Sai Bharadhwaj Matha +3Aerial Robotics Safe Robot Navigation
Fraunhofer Institute IVI · TU Munich · Munich Center for Machine Learning
Scaling Datasets for Multi-Sensor, Multi-Agent, and Multi-Domain Learning in Autonomous Systems
Jun 3, 2026R. Spencer Hallyburton, David Hunt, Miroslav PajicSynthetic Data Generation for Robotics Multi-Agent Systems
aDepartment of Electrical and Computer Engineering, Duke University, Durham, NC 27708 USA
A Multimodal Large Language Model-Driven Framework for Context-Aware UAV Emergency Landing Site Selection
Feb 1, 2026Chunliang Hua, Lei Zhang, Jiayang Sun +2UAV Navigation Multimodal Large Language Models
School of Information Science and Engineering, Southeast University, Nanjing 211189, China · LASER, International Digital Economy Academy, Shenzhen 510085, China · Department of Electronic Engineering, East China Normal University, Shanghai 200241, China

Mar 18, 2026J/K move · Enter open · S save

SafeLand: Safe Autonomous Landing in Unknown Environments with Bayesian Semantic Mapping

Markus Gross, Andreas Greiner, Sai Bharadhwaj Matha, Felix Soest+2

Fraunhofer Institute IVI · TU Munich · Munich Center for Machine Learning

Aerial Robotics Safe Robot Navigation Semantic Mapping

Open paper AskPDF arXiv

Autonomous landing of uncrewed aerial vehicles (UAVs) in unknown, dynamic environments poses significant safety challenges, particularly near people and infrastructure, as UAVs transition to routine urban and rural operations. Existing methods often rely on prior maps, heavy sensors like LiDAR, static markers, or fail to handle non-cooperative dynamic obstacles like humans, limiting generalization and real-time performance. To address these challenges, we introduce SafeLand, a lean, vision-based system for safe autonomous landing (SAL) that requires no prior information and operates only with a camera and a lightweight height sensor. Our approach constructs an online semantic ground map via deep learning-based semantic segmentation, optimized for embedded deployment and trained on a consolidation of seven curated public aerial datasets (achieving 70.22% mIoU across 20 classes), which is further refined through Bayesian probabilistic filtering with temporal semantic decay to robustly identify metric-scale landing spots. A behavior tree then governs adaptive landing, iteratively validates the spot, and reacts in real time to dynamic obstacles by pausing, climbing, or rerouting to alternative spots, maximizing human safety. We extensively evaluate our method in 200 simulations and 60 end-to-end field tests across industrial, urban, and rural environments at altitudes up to 100m, demonstrating zero false negatives for human detection. Compared to the state of the art, SafeLand achieves sub-second response latency, substantially lower than previous methods, while maintaining a superior success rate of 95%. To facilitate further research in aerial robotics, we release SafeLand's segmentation model as a plug-and-play ROS package, available at https://github.com/markus-42/SafeLand.

Abstract

Aerial Robotics

Explore similar work

SafeLand: Safe Autonomous Landing in Unknown Environments with Bayesian Semantic Mapping

Scaling Datasets for Multi-Sensor, Multi-Agent, and Multi-Domain Learning in Autonomous Systems

A Multimodal Large Language Model-Driven Framework for Context-Aware UAV Emergency Landing Site Selection