Corina Iovan

Significant progress has been made in recent years concerning the automatic reconstruction of man-made objects or environments from multiple aerial images. Yet, a lot of challenge concerning the modelling of other objects present on the terrain surface, such as trees, shrubs, hedges, or lawns still exists. An accurate reconstruction of such types of vegetation areas is a challenge due to their complex nature and to their intricate distribution between man-made objects in dense urban areas. This thesis presents an image analysis system for vegetation detection and characterisation from high resolution colour infrared aerial imagery for 3D city modelling. The aim of the system's first module is to extract vegetation areas. The approach developed is based on a Support Vector Machines (SVM) classifier and its performances are compared to traditional remote sensing methods for vegetation detection. The system's following modules aim at characterising vegetation areas thus identified, according to their morphology. Separation into high- (tree) and low- (lawn) height vegetation areas is based on texture characteristics computed on the digital surface model (DSM). Individual tree crown delineation is performed by using a region-growing algorithm based on geometrical characteristics of trees. 3D morphological characteristics (height, crown diameter, tree trunk position) are estimated for each tree crown. A supervised classification to characterise each tree by its species is performed on each tree crown. The set of parameters extracted by each of the modules are used to enrich 3D city models by virtual realistic tree models.

title = {Detection and characterisation of vegetation in urban areas from high-resolution aerial imagery},
author = {Corina Iovan},
school = {Université Pierre et Marie Curie - Paris 6},
year = {2009},
month = {Nov. }

Due to its complexity, vegetation analysis and reconstruction from remote sensing data is a challenging problem. Using architectural tree models combined with model inputs estimated from aerial image analysis, this paper presents an analysis-synthesis approach for urban vegetation detection, modelling and reconstruction. Tree species, height and crown size information are extracted by aerial image analysis. These variables serve for model inversion to retrieve plant age, climatic growth conditions and competition with neighbours. Functional-structural individual-based tree models are used to reconstruct and visualize virtual trees and their time evolutions realistically in a 3D viewer rendering the models with geographical coordinates in the reconstructed scene. Our main contributions are: (i) a novel approach for generating plant models in 3D reconstructed scenes based on the analysis of the geometric properties of the data, and (ii) a modelling workflow for the reconstruction and growth simulation of vegetation in urban or natural environments.

title = {Model-Based Analysis-Synthesis for Realistic Tree Reconstruction and Growth Simulation},
author = {Iovan, C., Cournède, P-H, Guyard, T., Bayol, B., Boldo, D. and Cord, M.},
journal = {IEEE Transactions on Geoscience and Remote Sensing (TGRS)},
year = {2013},
month = {March}

This paper presents a hierarchical color infrared aerial image analysis system for urban vegetation detection and characterization for 3D city modeling applications. The developed process starts by a vegetation area detection step. We present an approach based on a supervised classification method using Support Vector Machines (SVM) classifier which we will compare to traditional remote sensing methods based on spectral indexes. Vegetation areas identified are further on characterized according to species and morphology. Distinction between hi-height (trees) and low-height (lawns) vegetation areas is performed by a texture criteria computed on the Digital Surface Model (DSM). ). Tree crown borders are identified through a region growing algorithm based on tree-shape parameters. An SVM classifier gives the species class for each tree-region thus identified. This classification is used to enhance the appearance of 3D city models by a realistic representation of vegetation according to the vegetation land use, shape and tree species.

title = {Modélisation de la végétation en milieu urbain : détection et caractérisation à partir d'images aériennes haute résolution couleur et infrarouge},
author = {Iovan, C. and Boldo, D. and Cord, M.},
journal = {Revue Française de Photogrammétrie et de Télédétection},
year = {2009},
volume = {189},
pages = {17-27},
month = {March}

Research in the area of 3-D city modeling from remote sensed data greatly developed in recent years with an emphasis on systems dealing with the detection and representation of man-made objects, such as buildings and streets. While these systems produce accurate representations of urban environments, they ignore information about the vegetation component of a city. This paper presents a complete image analysis system which, from high-resolution color infrared (CIR) digital images, and a Digital Surface Model (DSM), extracts, segments, and classifies vegetation in high density urban areas, with very high reliability. The process starts with the extraction of all vegetation areas using a supervised classification system based on a Support Vector Machines (SVM) classifier. The result of this first step is further on used to separate trees from lawns using texture criteria computed on the DSM. Tree crown borders are identified through a robust region growing algorithm based on tree-shape criteria. A SVM classifier gives the species class for each tree-region previously identified. This classification is used to enhance the appearance of 3-D city models by a realistic representation of vegetation according to the vegetation land use, shape and tree species.

title = {Detection, Characterization and Modeling Vegetation in Urban Areas from High Resolution Aerial Imagery},
author = {Iovan, C. and Boldo, D. and Cord, M.},
journal = {IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing},
year = {2008},
volume = {1},
pages = {206-213},
number = {3},
month = {September}

title = {Moving and calling: Mobile phone data quality measurements and spatiotemporal uncertainty in human mobility studies},
author = {Iovan, C., Olteanu-Raimond, A.-M., Couronné, T., and Smoreda, Z.},
booktitle = {16th International Conference on Geographic Information Science (AGILE'13)},
year = {2013},
editor = {Springer},
month = {May}

This paper addresses the challenging problem of scene classification in street-view georeferenced images of urban environments. More precisely, the goal of this task is semantic image classification, consisting in predicting in a given image, the presence or absence of a pre-defined class (e.g. shops, vegetation, etc.). The approach is based on the BOSSA representation, which enriches the Bag of Words (BoW) model, in conjunction with the Spatial Pyramid Matching scheme and kernel-based machine learning techniques. The proposed method handles problems that arise in large scale urban environments due to acquisition conditions (static and dynamic objects/pedestrians) combined with the continuous acquisition of data along the vehicle’s direction, the varying light conditions and strong occlusions (due to the presence of trees, traffic signs, cars, etc.) giving rise to high intra-class variability. Experiments were conducted on a large dataset of high resolution images collected from two main avenues from the 12th district in Paris and the approach shows promising results.

title = {Classification of Urban Scenes from Geo-referenced Images in Urban Street-View Context},
author = {Iovan C., Picard D., Thome N., Cord M.},
booktitle = {Machine Learning and Applications (ICMLA), 2012 11th International Conference on},
year = {2012},
editor = {IEEE},
volume = {2},
pages = {339 -344},
month = {Dec}

An approach for tree species classification in urban areas from high resolution colour infrared (CIR) aerial images and the corresponding Digital Surface Model (DSM) is described in this paper. The proposed method is a supervised classification one based on a Support Vector Machines (SVM) classifier. Texture features from the Gray Level Co-occurrence Matrix (GLCM) are computed to form feature vectors for both per-pixel and per-region classification approaches. The two approaches are presented and results obtained are evaluated and compared both against each other and also against a manual defined ground truth. To perform tree species classification on high- density urban area images, trees must previously be segmented into individual objects. All intermediary methods developed to segment individual trees will also be shortly described. Tree parameters (height, crown diameter) are estimated from the DSM. These parameters together with the tree species information are used for a 3D realistic modelling of the trees in urban environments. Results of the described system are presented for a typical scene.

title = {Detection, segmentation and characterization of vegetation in high-resolution aerial images for 3D city modeling},
author = {Iovan, C. and Boldo, D. and Cord, M.},
booktitle = {International Archives of Photogrammetry, Remote Sensing and Spatial Information Sciences},
year = {2008},
editor = {ISPRS},
volume = {37},
number = {Part 3A},
month = {July}

This paper presents a complete high resolution aerial-images processing workflow to detect and characterize vegetation structures in high density urban areas. We present a hierarchical strategy to extract, analyze and delineate vegetation areas according to their height. To detect urban vegetation areas, we develop two methods, one using spectral indices and the second one based on a Support Vector Machines (SVM) classifier. Once vegetation areas detected, we differentiate lawns from treed areas by computing a texture operator on the Digital Surface Model (DSM). A robust region growing method based on the DSM is proposed for an accurate delineation of tree crowns. Delineation results are compared to results obtained by a Random Walk region growing technique for tree crown delineation. We evaluate the accuracy of the tree crown delineation results to a reference manual delineation. Results obtained are discussed and the influential factors are put forward.

title = {Automatic Extraction and Classification of Vegetation Areas from High Resolution Images in Urban Areas},
author = {Iovan, C. and Boldo, D. and Cord, M. and Erikson, M.},
booktitle = {Proc. of the 15th Scandinavian Conference on Image Analysis (SCIA)},
year = {2007},
editor = {Bjarne K. Ersboll and Kim Steenstrup Pedersen},
volume = {4522},
series = {Lecture Notes in Computer Science},
pages = {858--867},
address = {Aalborg, Denmark},
month = {Juin},
publisher = {Springer}

This paper presents a method for automatic extraction and characterisation of vegetation structures (such as trees, shrubs, hedges or lawns) in high density urban areas. We present a hierarchical strategy to extract, analyze and delineate vegetation areas according to their height. Spectral indices are used to detect urban vegetation areas. We differentiate lawns from treed areas by computing a texture operator on the digital elevation model (DEM) corresponding to the vegetation areas previously detected. A robust region growing method based on the DEM is developed for an accurate delineation of tree crowns. We evaluate the accuracy of the tree crown delineation results to a reference manual delineation. Results obtained are discussed and the influential factors are put forward.

title = {Automatic Extraction of Urban Vegetation Structures from High Resolution Imagery and Digital Elevation Model},
author = {Iovan, C. and Boldo, D. and Cord, M.},
booktitle = {Urban Remote Sensing Joint Event},
year = {2007},
pages = {1--5},
address = {Paris, France},
month = {Avril}

Nous présentons un système hiérarchique pour analyser la végétation urbaine à partir d'images aériennes couleur et infrarouges à très haute résolution et d'un modèle numérique d'élévation. Le système proposé détecte les zones de végétation et les classifie en végétation haute et basse. Les houppiers de chaque arbre sont individualisés par un algorithme de segmentation d'images. Des paramètres géométriques (la hauteur, le diamètre de la couronne, la localisation du tronc des arbres) sont estimés pour chaque arbre. Une classification supervisée pour caractériser les espèces d'arbres est ensuite effectuée pour chaque objet ainsi extrait. Finalement, un rendu réaliste de l'environnent virtuel urbain est obtenu en intégrant des modèles virtuels d'arbres.

title = {Détection et caractérisation automatique de la végétation à partir d'images aériennes très haute résolution},
author = {Iovan, C.},
booktitle = {Bulletin d'information scientifique et technique de l'IGN},
year = {2011},
address = {Saint-Mandé, France},
volume = {78},
month = {May},
organization = {Institut Géographique National (IGN)}

The Pleiades system developed by the French space agency CNES shall be ready for launch in early 2009. It deals with a new generation of high resolution optical satellites. The great agility of Pleiades satellites with their motion capability in roll, pitch, and heading, will allow new acquisition modes which are impossible with SPOT series satellites. Indeed, the system will be able to perform three-fold stereoscopic colour images of 20km swath and 70cm resolution. Several experimentations and simulations are being carried out in order to verify the ability of the system to fulfill the technical and operational requirements of the map-making processes. Last year, a first study has been carried out on 3D city models production on two small areas from simulated images and with a lot of different contexts. Considering the promising results, studies carry on this year, with more different areas and Pleiades simulation images more realistic. Two other topics have been studied : automatic vegetation detection and DTM processing. Concerning building extraction we have used the same production line as last year. Concerning vegetation extraction we tested two methods. The first one is classic : we threshold the Normalized Difference Vegetation Index (NDVI). The second method is more elaborated : it deals with an algorithm which studies the value of texture features and transforms colour spaces for segmentation and classification. For the DTM calculation we have used a method based on elastic grid algorithm. Finally we have integrated all these different necessary primitives (buildings, vegetation and DTM) to produce a complete 3D model on new datatests areas.

title = {Evaluation of the potential of PLEIADES system for 3D city models production: building, vegetation and DTM extraction},
author = {Durupt, M. and Flamanc, D. and Le$-$Bris, A. and Iovan, C. and Champion, C.},
booktitle = {Proc. of Commission I Symposium},
year = {2006},
address = {Marne-la-Vallée, France},
volume = {36},
number = {Part 1},
month = {July},
organization = {International Society for Photogrammetry and Remote Sensing (ISPRS)}

title = {Study and Emulation of the Internet Banking Service},
author = {Iovan, C. and Borda, M. and Amadou, K.},
booktitle = {Proceedings of Conference of Energetics and Electrical Engineering (ENELKO)},
year = {2004},
pages = {97--103},
address = {Budapest, Hungary}

title = {A Secure and Robust Image Digital Watermarking},
author = {Major, R. and Borda, M. and Iovan, C.},
booktitle = {COST 276 Workshop},
year = {2004},
address = {Budapest, Hungary}