The Change You Want To Detect: Semantic Change Detection In Earth Observation With Hybrid Data Generation

Yanis Benidir, Nicolas Gonthier, Clément Mallet · 2025

Read The Change You Want To Detect: Semantic Change Detection In Earth Observation With Hybrid Data Generation by Yanis Benidir, Nicolas Gonthier, Clément Mallet (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2503-15683.txt.

Introduction

Yanis Benidir, Nicolas Gonthier, and Clément Mallet address change detection in satellite imagery, in a two thousand twenty-five paper. Comparing very high resolution images from two dates can reveal land cover change. Semantic methods need large annotated datasets, while simpler binary methods cover only narrow cases.

The authors present HySCDG, a generative pipeline that mixes real images with inpainted ones. It produces land cover maps for both dates plus the change map. The resulting dataset, FSC-180k, is large, realistic, and diverse. They test it on five change detection setups, from zero-shot to low-data training. Pretraining on the hybrid data significantly boosted performance. The approach aims at systems that adapt across places and times.

Catalog record

CREATORYanis Benidir, Nicolas Gonthier, Clément Mallet
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE57 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2503-15683

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

The Change You Want To Detect: Semantic Change Detection In Earth Observation With Hybrid Data Generation Yanis Benidir, Nicolas Gonthier, Clément Mallet Abstract Bi-temporal change detection at scale based on Very High Resolution (VHR) images is crucial for Earth monitoring. This remains poorly addressed so far: methods either require large volumes of annotated data (semantic case), or are limited to restricted datasets (binary set-ups). Most approaches do not exhibit the versatility required for temporal and spatial adaptation: simplicity in architecture design and pretraining on realistic and comprehensive datasets. Synthetic datasets are the key solution but still fail to handle complex and diverse scenes. In this paper, we present HySCDG a generative pipeline for creating a large hybrid semantic change detection dataset that contains both real VHR images and inpainted ones, along with land cover semantic map at both dates and the change map. Being semantically and spatially guided, HySCDG generates realistic images, leading to a comprehensive and hybrid transfer-proof dataset FSC-180k. We evaluate FSC-180k on five change detection cases (binary and semantic), from zero-shot to mixed and sequential training, and also under low data regime training. Experiments demonstrate that pretraining on our hybrid dataset leads to a significant performance boost, outperforming SyntheWorld, a fully synthetic dataset, in every configuration. All codes, models, and data are available here: https://yb23.github.io/projects/cywd/ . 1 Introduction Figure 1 : Efficient and scalable change detection requires a comprehensive training dataset that does not exist today. Using a single-temporal dataset (image+semantic map), we propose HySCDG that generates a novel bi-temporal hybrid dataset FSC-180k. This enables multiple transfer learning possibilities on either binary and semantic change detection tasks. / : Transfer with/without fine-tuning our model, respectively. Efficient methods for change detection (CD) are crucial for monitoring territories and the various phenomena and activities that impact them Burke_Science . As human and climatic pressures increase, this topic is of utmost importance and calls every year for higher spatial resolution and semantic accuracy Nature_swin . Change detection at scale has benefited greatly from the contributions of deep learning in recent years daudt2018fully ; ZHENG2024239 ; zheng2024changen2 ; ChangeMamba , but most methods are grounded on large volumes of annotated data Van_Etten_2021_CVPR ; Priya_fire . Creating a large-scale dataset for bi-temporal (pair-wise) remote sensing change detection poses significant challenges and costs, especially when dealing with Very High Resolution (VHR) images (ground sampling distance: GSD < 1 m, Li_2024_CVPR ). This is due to the need for expertise and considerable effort in collecting, preprocessing, and annotating images Toker_2022_CVPR ; xia2023openearthmap , particularly in the case of VHR data and multiple labels (termed semantic change detection , SCD) Zheng_2021_ICCV ; ZHU2022113266 ; TIAN2022164 ; WANG2024103761 . As an alternative, synthetic data generation is a promising direction for bi-temporal labeled data provision from a single temporal domain that is successful in diverse areas such as urban sprawl SMARS , disaster or climate change monitoring 9420150 or crop mapping ijgi12110450 . Two main paradigms exist for CD: fully synthetic datasets (mainly graphics-based kolos2019procedural ; SyntheWorld ), and hybrid solutions that mix real and modified images, often at the instance level chen2021adversarial ; Changen ; zheng2024changen2 , known as data augmentation . The first method synthesizes image pairs by rendering two versions of a 3D virtual world thanks to a computer graphics engine. On the other hand, hybrid approaches synthesize new image pairs by pasting or erasing object instances into existing images ( inpainting ). They propose a suitable trade-off and faster generation of large quantities of data, with likely higher realism. However, none of these solutions meet the requirements for scalable SCD with VHR imagery ( Table 1 ): focus on single tasks, restricted geographical areas, limited diversity, no guarantee on semantic consistency between pairs SEO_wacv and subsequently poor transfer to real datasets. To address this challenge, we introduce a generative pipeline built upon Stable Diffusion LatentDiffusionModels and ControlNet ControlNet , designed to harness an existing VHR semantic land cover dataset garioud2023flair and instance masks, to generate an extensive hybrid semantic bi-temporal change detection dataset encompassing both real and inpainted images, semantic land cover maps at each date, and a change map ( Fig. 1 ). Then, we evaluate the transferability of our synthetic dataset on five datasets for both binary and semantic change detection cases ranging from zero-shot to sequential and mixed training set-ups, as well as under low-data regime. Our contributions are as follows: • We present a generation pipeline HySCDG for creating VHR bi-temporal SCD datasets based on a land cover real-world dataset by using Stable Diffusion and ControlNet with tunable and semantically-guided inpainting of individual objects. It can be easily configured or finetuned on any land-cover dataset, generating images with the dataset’s distinctive traits. • We provide FSC-180k : a Hybrid Semantic Change Detection pretraining dataset , created from FLAIR garioud2023flair along with the instance masks for 300k objects for the corresponding images. • We provide a comprehensive evaluation of the proposed synthetic dataset on multiple transfer learning scenarios on five change datasets, in particular for SCD. 2 Related work Detecting changes in the deep learning era. We target bi-temporal land cover change detection in VHR images lv2021land . This historical task howarth1981procedures ; singh1989review ; nemmour2006multiple consists in detecting the meaningful elements of a given area that have changed between two acquisitions and providing the land cover categories at each period (also known as change trajectory ). It presents challenges: radiometric variations, geometric variations and non permanent changes. They have been extensively addressed with supervised deep learning techniques Zheng_2021_ICCV : from CNN daudt2018fully ; zhan2017change ; daudt2018hrscd to Transformers ZHENG2022228 ; ZHENG2024239 ; ChangeMamba ; ChangeFormer ; SCanNet ; Nature_swin , mainly with Siamese architecture. Multiple datasets, accompanied with ad-hoc methods, have been created to feed such models. They are often limited by their size hiucd ; LevirCD ; SECOND ; pang2023detecting ; ji2018fully_WHU , geographical extent LevirCD ; SECOND ; hiucd ; TIAN2022164 , low geometrical or annotation quality daudt2018hrscd . Also, most of the effort focuses on binary change detection, i.e. , restricted to a unique category, mainly buildings LevirCD ; s2looking ; pang2023detecting ; ji2018fully_WHU ; peng2020semicdnet or disasters ZHANG20231 . Such scarcity in high quality, diverse, large-scale labels is alleviated with transfer learning shi2020change ; CAO2023113371 , by fine-tuning models pretrained on large-scale datasets for different image domains, even for semantic change detection SECOND ; hiucd ; TIAN2022164 . Synthesizing and inpainting remote sensing images. Synthetic remote sensing data is useful e.g., for cloud removal czerkawski2022deep , image restoration fan2011pixel , or training deep supervised models toker2024satsynth , before transfer to real examples. Solutions based on patches, meng2017sparse ; zheng2021nonlocal , autoencoders huang2022image , GAN marin2021sssgan ; kuznetsov2020remote , pixel-aligned generation isola2017image have been proposed, but with limited image quality and no semantic control. Diffusion models (DM) have improved such quality LatentDiffusionModels , but most models have applied DM directly on mid-resolution RGB images khanna2023diffusionsat ; GenerateYourOwnScotland ; toker2024satsynth , neglecting both the multi-spectral dimension czerkawski2024exploring and VHR spaces. However, a key finding remains: semantically control the synthesis, e.g. , by directly synthesizing text khanna2023diffusionsat , semantic maps toker2024satsynth or with a control module guided by edges czerkawski2024exploring , semantic maps GenerateYourOwnScotland or metadata khanna2023diffusionsat .

Read the complete text