Water is an indispensable natural resource for life. Despite the fact that 72% of the Earth's surface is covered with water, only 3% of these water reserves are in the form of fresh water, 96% of which constitutes ice-poles. As a result, countries such as Canada with nearly 9% of the world's renewable water resources may be considered a freshwater-rich country, hence the importance of water bodies mapping as an essential tool for water resources inventory and monitoring. This research aims at developing an efficient method for river extraction from Synthetic Aperture Radar (SAR) images. Our proposed approach adapted for SAR images is mainly based on multi-criteria image filtering using attribute filters in order to identify water bodies and distinguish rivers from other structures present in the input image. First a texture measurement is applied to the input image in order to extract homogeneous areas, then a hierarchical image representation based on the Max-Tree structure is used to filter objects according to their shape and size. Our technique has been tested on different study sites acquired using RADARSAT-2 and Sentinel-1 satellites. Images used in the experimentation represent urban and rural areas characterized by rivers of variable dimensions.

Floods rank among the most devastating and deadly disasters likely to occur. Each year, thousands of people die as a result of rising river water levels and spring snow melts. Minimizing the impact of these events and providing necessary information to rescue teams on the ground requires the mobilization of all humanitarian stakeholders. This work aims to automate the extraction of flooded areas from Synthetic Aperture Radar (SAR) satellite images in order to accelerate the generation of damage maps during floods. Our approach is mainly based on the following steps: first, homogeneous surfaces contained in pre- and post-disaster images are extracted based on the Structural Feature Set (SFS) texture descriptor. Then, a morphological opening operator is applied to the resulting images in order to filter out small-sized objects and noise. Finally, the two gray levels images generated by applying the texture descriptor and the morphological operator are fused using fuzzy logic to identify flooded regions. SAR images acquired using the RADARSAT-2 satellite during the Richelieu River floods are used to evaluate the performance of the proposed technique, and the results obtained show the efficiency and robustness of the described approach.

Water surface extraction using satellite images proves to be of great importance due to its utility in several applications such as land use, floods management and monitoring. Among the wide range of sensors orbiting around the earth, Synthetic Aperture Radar (SAR) proves to be a very effective tool in this context due to its robustness to unfavorable weather conditions and its cloud penetrating capabilities. This paper presents a novel rivers extraction method from SAR images mainly based on the combination of a local texture measurement and global knowledge associated to the shape of the object of interest. A local texture measurement is first computed for every pixel of the image to extract homogeneous surfaces, then a mathematical morphology operator is applied to attenuate noise generated by speckle characterizing SAR images. Finally, the surface occupied by the object of interest is compared to the surface associated to the smallest rectangle that encloses this object in order to separate rivers from lakes in the image. The proposed approach was tested on SAR images acquired by RADARSAT-2 satellite from numerous regions of Canada. Our experimental results demonstrate that the proposed approach is robust and effective.

Each year, numerous disasters cause high amount of human and material losses. With the presence of both technological and social means like very high spatial resolution (VHR) satellite images and the International Charter “Space and Major Disasters” respectively, decision makers can obtain the needed information to make fast life-saving decisions. The automation of parts of the image analysis process is thus of great interest. This is why we propose to apply an ensemble classification method to provide better building damage evaluation using optical images acquired before and after the event. We base this work on our previous framework for fast building detection and damage evaluation by supervised classification. The results show the positive impact of ensemble classifiers in damage mapping.

Infrastructures damage detection in case of major disasters is one of the most discussed problems and represent an active field of research in remotely sensed imaging. In this paper, a novel method designed for fast roads damage extraction is proposed since these structures are important in the delivery of assistance and to manage the intervention of the emergency teams on ground. The proposed methodology includes first an already completed step that consists in extracting the road network from both the pre- and post-disaster images. Then, a multiscale segmentation based on the wavelet transform is performed on the road surface and the obtained objects from the two coregistered images are compared. Finally, the Dempster Shafer theory is applied to decide the membership class of each object in a first step, and then identify the nature of changes using the multidimensional evidential reasoning. Images acquired by the Geo-Eye satellite before and after the earthquake that hits Port-au-Prince (Haiti) on January 2010 are used in the experiments.

The International Charter1"Space and Major Disasters" is regularly activated during a catastrophic event and offers rescue teams comprehensive damage maps. Most of these maps are built by means of satellite image manual processing, which is often complex and demanding in terms of time and energy. Automatic processing supplies prompt treatment. Nevertheless, it usually presents a semantic gap handicap. The exploitation of ontologies to bridge the semantic gap has been widely recommended due to their quality of knowledge representation, expression, and discovery. In this work, we present an ontology-based semantic hierarchical classification method to undertake this problem. Ontology components are translated to image-based parameters and exploited to assist the classification process at two levels, and using 12 classes. The region of interest is selected from the first level and exhaustively analyzed and classified at the second level. The 2010 Haiti earthquake was selected as study area for this work. Experiments were performed using very high resolution multi-temporal QuickBird imagery and eCognition software.

When a disaster occurs, earth observation specialists must analyze a considerable amount of images as quickly as possible to help rescue teams or to make on-the-fly decisions. As there is currently no remote sensing software that is specific for natural disasters, we propose in this work to determine the basic requirements for such software and develop a test framework for such application. Our approach includes a state of the art review on current earth observation software as well as disaster response needs. We also list requirements based on our review and propose a solution for providing software assistance to photointerpreters. Finally our test framework is presented.

Road extraction is a topical research because of complexity due to his large topological variability. Increasing the spatial resolution generates noise which makes extraction difficult, especially in case of major disaster in an urban context. This problem increases false alarm rates and generally affects the performance of road extraction algorithm. Our aim is to improve the quality of roads extraction after adaptation of the Lowe's SIFT descriptors (scale-invariant feature transform) jointly with spectral angle algorithm. The characterization is performed on two image at various resolution images, respectively representing a rural and urban disaster area, captured by Quickbird satellite. Our approach significantly reduces the amount of false detection and shows an overall accuracy of up to nearly 30% in some cases.

In this paper, a new method for road extraction from very high resolution satellite images is introduced. The proposed methodology is based on multiscale analysis of the image through beamlet transform that can approximate edges of an image by means of line segments appearing at different resolution scales. The hypothesis of the membership of these segments to a road is verified by identification of parallel beam-lets representing the main axes of the road. A reconstruction step of the road network is finally applied by aggregation of beamlets representing road candidates belonging to lower scales.

Building damage evaluation is an important part of disaster response and recovery phases of the emergency management cycle. Earth observation data acquired by remote sensing satellites can provide useful information for damage mapping. Unfortunately, very high spatial resolution images are quite large and require analysis by multiple expert photo-interpreters in order to extract meaningful information in short time. This can lead to human errors caused by fatigue and varying degrees of image understanding from one analyst to the next. In this paper, we propose an object-based supervised learning scheme using geometric, scale and textural object features to detect the level of damage on buildings affected by an earthquake. The method is applied to the case of the 2010 Haiti earthquake.

Natural and man-made disasters make the headline many times a year. Crisis images depicting extent of event and intensity of damage are frequently displayed to emphasize the impact of the disaster on the local population. These maps are more and more often generated from raw images acquired by spaceborne sensors. Very high space resolution images are huge and require very careful analysis by numerous teams of trained photo-interpreter in order to provide meaningful maps in a timely manner. This paper describes the acquisition process and image analysis steps currently required to create crisis maps for events for which the International Charter “Space & Major Disasters” is activated. We answer the question “how can image processing time be reduced for faster disaster mapping?” by proposing a semi-automated building detection process.

The need for fast and reliable remote sensing algorithms is continuously growing. Remote sensing libraries are scarce and sometimes are difficult to use. The Orfeo Toolbox (OTB) is one such library that stands apart with its robust development scheme and programming concepts. Not all remote sensing researchers have a strong programming background. Programmers need to find ways of giving easier access to desired algorithms without the need of cumbersome programming overheard. This paper compares the actual parameter management class of OTB with a proposed dynamic graphical interface to replace it. The programming concepts used will first be presented. This work will lead to creating a visual programming environment for OTB.

We propose an automated inspection system of manufactured parts using a cloud of 3D measured points of a part provided by a range sensor, and its CAD model. Inspection consists in verifying the accuracy of a part related to a given set of tolerances. It is thus necessary that the 3D measurements be accurate. In the 3D capture of a part, several sources of error can alter the measured values. So, we have to find and model the most influential parameters affecting the accuracy of the range sensor in the digitalization process. This model is used to produce a sensing plan to acquire accurately the geometry of a part. By using the noise model, we introduce a dispersion value for each 3D point acquired. This value of dispersion is shown as a weight factor in the inspection results.

The analysis of natural linear structures, termed "lineaments" in satellite images, provides important information to the geologist. In the satellite imaging process, important features of the observed tridimensional scene, including geological lineaments, are mapped into the resulting 2D image as sharp radiation variations or edge elements (edgels). Edgels are detected by a first-order differentiation operator and are linked together with those in the vicinity on a basis of orientation continuity. Lineaments are mapped into remotely sensed satellite images as long and continuous quasilinear features and can be described as a connected sequence of edgels whose direction may change gradually along the sequence. Parts of the same lineament can be occluded by geomorphological features and must be linked together, a major drawback with local and small neighborhood detectors. The authors propose a cellular neural network (CNN) architecture to offer a large directional neighborhood to the lineament detection algorithm. The CNN uses a large circular neighborhood coupled with a directional-induced gradient field to link together edgels with similar and continuous orientation. Missing edgels are restored if a surrounding lineament is detected.

Online inspection for defects in manufactured parts is a challenging problem and a crucial issue for any production process. Precise and extended CAD data describe the manufactured part under inspection and can be stored in a database in order to perform comparisons between the measured observed part (with a 3D laser camera, for instance) and its CAD description. The problem is very slow access time to find the right CAD model, because of the huge amount of data necessary to represent even a very simple manufactured piece. We propose a fast access to the database based on a vision-derived representation of the part, much less precise than the CAD representation but which offers a very low storage requirement. CAD representations of the models are converted off-line into a vision-based representation used to train a neural network. Volumetric projected components (geons) and their interrelationship are extracted from the digitized image of the piece under inspection and input to a neural network whose outputs point to the most probable models in the large database. Corresponding CAD data is then available for the inspection module.

A large number of computer programs are discussed in the scientific literature for the implementation of most digital signal processing algorithms. The reader interested in evaluating this type of algorithm is often discouraged by the time and effort he has to spend interfacing the algorithm to his particular processing environment. This paper presents a software testbed package called the Digital Signal Processing Interactive System (DSPIS) which was designed to be a unified and consistent environment for the evaluation of digital signal processing algorithms. Input and output data are confined to seven vectors: X, Y and V for time domain, A, B and W for spectral domain and, finally, C for coefficient domain. Data can be synthesized or transferred from a file into one of the vectors, processed by one or more algorithms and the results (in one or many vectors) can be stored in files. Any vector can be printed out or displayed against its scaled index (except the vector C). DSPIS is menu oriented and permits interactive execution of commands.