Search

International audience; A general framework of spatio-spectral segmentation for multi-spectral images is introduced in this paper. The method is based on classification-driven stochastic watershed (WS) by Monte Carlo simulations, and it gives more regular and reliable contours than standard WS. The present approach is decomposed into several sequential steps. First, a dimensionality-reduction stage is performed using the factor-correspondence analysis method. In this context, a new way to select the factor axes (eigenvectors) according to their spatial information is introduced. Then, a spectral classification produces a spectral pre-segmentation of the image. Subsequently, a probability density function (pdf) of contours containing spatial and spectral information is estimated by simulation using a stochastic WS approach driven by the spectral classification. The pdf of the contours is finally segmented by a WS controlled by markers from a regularization of the initial classification.

This effort we acquired and registered a multi-spectral dynamic image test set with the intent of using the imagery to assess the operational effectiveness of static and dynamic image fusion techniques for a range of relevant military tasks. This paper describes the image acquisition methodology, the planned human visual performance task approach, the lessons learned during image acquisition and the plans for a future, improved image set, resolution assessment methodology and human visual performance task. © 2009 SPIE.

This effort we acquired and registered a multi-spectral dynamic image test set with the intent of using the imagery to assess the operational effectiveness of static and dynamic image fusion techniques for a range of relevant military tasks. This paper describes the image acquisition methodology, the planned human visual performance task approach, the lessons learned during image acquisition and the plans for a future, improved image set, resolution assessment methodology and human visual performance task. © 2009 SPIE.

The Atmospheric-Space Interactions Monitor (ASIM) consists of a suite of instruments to make observations of Transient Luminous Events (TLEs) and Terrestrial Gamma-ray Flashes (TGFs). The optical part of the payload is called the Multi-Spectral Modular Imaging Array (MMIA) and consists of a combination of a camera units and photometers. The Modular X- and Gamma-ray Sensor (MXGS) is a coded aperture based gamma ray imaging system. The payload is to be own on the ISS to measure the TLE's which are associated with electrical discharges in the atmosphere above 10km and commonly associated with lightning events on the Earth. Three primary types of TLE's will be studied: SPRITES, ELVES and HALOES. There will be six photometers measuring photon fluxes in specific narrow band and broad band channels between 145nm and 740nm. The six photometers will be mounted into three MMIA units each containing two photometers. Two MMIA modules will be pointed towards the Earth Limb and on MMIA to the Nadir. This paper describes the preliminary design of the photometers based on the current measurements and models of TLE characteristics.

We present an algorithm that produces a fused false color representation of a combined multiband IR and visual imaging system for maritime applications. Multispectral IR imaging techniques are increasingly deployed in maritime operations, to detect floating mines or to find small dinghies and swimmers during search and rescue operations. However, maritime backgrounds usually contain a large amount of clutter that severely hampers the detection of small targets. Our new algorithm deploys the correlation between the target signatures in two different IR frequency bands (3-5 and 8-12 ìm) to construct a fused IR image with a reduced amount of clutter. The fused IR image is then combined with a visual image in a false color RGB representation for display to a human operator. The algorithm works as follows. First, both individual IR bands are filtered with a morphological opening top-hat transform to extract small details. Second, a common image is extracted from the two filtered IR bands, and assigned to the red channel of an RGB image. Regions of interest that appear in both IR bands remain in this common image, while most uncorrelated noise details are filtered out. Third, the visual band is assigned to the green channel and, after multiplication with a constant (typically 1.6) also to the blue channel. Fourth, the brightness and colors of this intermediate false color image are renormalized by adjusting its first order statistics to those of a representative reference scene. The result of these four steps is a fused color image, with naturalistic colors (bluish sky and grayish water), in which small targets are clearly visible. Keywords: target detection; false color fusion, small targets, clutter, image fusion

We present an algorithm that produces a fused false color representation of a combined multiband IR and visual imaging system for maritime applications. Multispectral IR imaging techniques are increasingly deployed in maritime operations, to detect floating mines or to find small dinghies and swimmers during search and rescue operations. However, maritime backgrounds usually contain a large amount of clutter that severely hampers the detection of small targets. Our new algorithm deploys the correlation between the target signatures in two different IR frequency bands (3-5 and 8-12 im) to construct a fused IR image with a reduced amount of clutter. The fused IR image is then combined with a visual image in a false color RGB representation for display to a human operator. The algorithm works as follows. First, both individual IR bands are filtered with a morphological opening top-hat transform to extract small details. Second, a common image is extracted from the two filtered IR bands, and assigned to the red channel of an RGB image. Regions of interest that appear in both IR bands remain in this common image, while most uncorrelated noise details are filtered out. Third, the visual band is assigned to the green channel and, after multiplication with a constant (typically 1.6) also to the blue channel. Fourth, the brightness and colors of this intermediate false color image are renormalized by adjusting its first order statistics to those of a representative reference scene. The result of these four steps is a fused color image, with naturalistic colors (bluish sky and grayish water), in which small targets are clearly visible. Keywords: target detection; false color fusion, small targets, clutter, image fusion

Automatic extraction of buildings and modeling of their spatial arrangements provide essential information for urban applications. This paper describes our work on modeling urbanization using spatial building patterns. Building detection is done using Bayesian classification of multi-spectral information. The individual buildings are used as textural primitives, and co-occurrence based spatial domain features and Fourier spectrum-based frequency domain features are used to model their repetitiveness and periodicity at particular orientations. These features are used to classify image neighborhoods as organized (regular) and unorganized (irregular). Experiments with high-resolution Ikonos imagery show that the proposed technique can be used for automatic segmentation of urban scenes and extraction of valuable information about urban growth.

Knowledge of water depth is a crucial for planning military amphibious operations. Bathymetry from remote sensing with multispectral or hyperspectral imagery provides an opportunity to acquire water depth data faster than traditional hydrographic survey methods without the need to deploy a hydrographic survey vessel. It also provides a means of collecting bathymetric data covertly. This research explores two techniques for deriving bathymetry and assesses them for use by those involved in providing support to military operations. To support this aim a fieldwork campaign was undertaken in May, 2000, in northern Queensland. The fieldwork collected various inherent and apparent water optical properties and was concurrent with airborne hyperspectral imagery collection, space-based multispectral imagery collection and a hydrographic survey. The water optical properties were used to characterise the water and to understand how they affect deriving bathymetry from imagery. The hydrographic data was used to assess the performance of the bathymetric techniques. Two methods for deriving bathymetry were trialled. One uses a ratio of subsurface irradiance reflectance at two wavelengths and then tunes the result with known water depths. The other inverts the radiative transfer equation utilising the optical properties of the water to derive water depth. Both techniques derived water depth down to approximately six to seven metres. At that point the Cowley Beach waters became optically deep. Sensitivity analysis of the inversion method found that it was most sensitive to errors in vertical attenuation Kd and to errors in transforming the imagery into subsurface irradiance reflectance, R(0-) units. Both techniques require a priori knowledge to derive depth and a more sophisticated approach would be required to determine water depth without prior knowledge of the area of interest. This research demonstrates that water depth can be accurately mapped with optical techniques in less than ideal optical conditions. It also demonstrates that the collection of inherent and apparent optical properties is important for validating remotely sensed imagery.

The recently approved ExoMars rover is the first element of the ESA Aurora programme and is scheduled to deliver the Pasteur exobiology payload to Mars by 2015. The 0.7 kg Panoramic Camera will provide multi-spectral stereo images with 65 ° field-of-view (1.1 mrad/pixel) and high-resolution (85 μrad/pixel) monoscopic ‘zoom’ images with 5 ° field-of-view. The stereo wide-angle cameras (WAC) are based on the Beagle 2 Stereo Camera System heritage (Griffiths et al. (2005). Planet. Space Sci. 53, 1466–1488). The Panoramic Camera instrument is designed to fulfil the digital terrain mapping requirements of the mission as well as to provide multi-spectral geological imaging, colour and stereo panoramic images, water vapour abundance and dust optical depth measurements. It can also be used for high-resolution imaging of inaccessible locations on crater walls and to observe retrieved subsurface samples before ingestion into the rest of the Pasteur payload.

The United States Department of Energy (DOE) is the lead federal agency for the research, development, demonstration, and deployment (RDD&D) of carbon storage technologies. Monitoring, verification, accounting and assessment (MVA) is an essential element of geologic CO 2 storage projects, since without MVA it is not possible to understand the fate of CO 2 in the injection formation or to monitor any potential CO 2 releases to underground sources of drinking water (USDW) or the atmosphere. Determining the location of CO 2 in the subsurface and identifying indicators of the potential release of CO 2 to the atmosphere are significant challenges requiring the adoption of existing technologies as well as novel approaches for monitoring large areas above a storage reservoir. Remote sensing technologies could offer a solution. Remote sensing refers to the use of monitoring tools that can gather data at a location remote to the area of interest, and could provide an option for non-invasive and large scale spatial monitoring. The National Energy Technology Laboratory (NETL) has been developing and deploying remote sensing applications over the past decade to improve monitoring of both geologic and terrestrial carbon storage projects. MVA remote sensing tools being developed or improved by NETL sponsored research include interferometric synthetic aperture radar (InSAR), tiltmeters/GPS, remote operated vehicles (ROVs), SEQURE™, light detection and ranging (LIDAR), and multi-spectral/hyper-spectral scanning. Benefits associated with NETL's remote sensing effort include development and deployment of non-intrusive tracking and monitoring technologies, reduced manpower requirements needed to meet potential regulated MVA requirements, and reduced costs to implement monitoring technologies. These technologies can provide early detection of CO 2 releases which, in turn, will allow for further refinement of project monitoring protocols using more conventional detection and mitigation technologies to better pinpoint CO 2 location. This paper addresses selected remote sensing techniques under NETL development and field deployment of these technologies.

Image registration aligns two images geometrically, which is frequently required in medical, computer vision and remote-sensing field. It is a crucial pre-processing step in change detection or growth monitoring using satellite images. Accuracy of change detection depends on accuracy of image registration. For multi-modal, multi-sensor, multi-spectral satellite images one of the challenges for image registration is varying illumination level according to the sensor characteristics. This challenge is addressed by using Histogram of Oriented Gradient (HOG) along with Speeded-Up Robust Feature (SURF). It is shown that illumination variation gives some incorrect matches with SURF only which degrades image registration. Incorrect matches are reduced by using HOG as descriptor in SURF. Supporting simulation results for satellite images are presented which show the improvement in the correct matching rate.

International audience; The aim of this paper is to study the influence of the measurement noise (on the flux) and of model bias on the temperature estimation by multispectral method with two, three or four colors using a model based on the Wien's approximation (and flux ratio) and taking into account spectral emissivity up to 2-order variations (a special case will be considered: variation according to the Drude's law). For this purpose, in a first part the model and the method will be presented, then in a second part, the estimation of parameters from a regularized nonlinear least squares regression algorithm (Levenberg-Marquardt), and their statistical properties obtained using a Monte Carlo method (1000 estimates) will be discussed. The study will conclude with experimental validations in the InfraRed range (using a thermal camera).

Industrial quality inspection using image analysis on astaxanthin coating in aquaculture feed pellets is of great importance for automatic production control. In this study multi-spectral image analysis of pellets was performed using LDA, QDA, SNV and PCA on pixel level and mean value of pixels for each pellet. Classication using LDA or QDA on pellet mean or median values showed better results than using the pixel values or PCA.

Infrastructure expands rapidly in the Arctic due to industrial development. At the same time, climate change impacts are pronounced in the Arctic. Ground temperatures are, for example, increasing as well as coastal erosion. A consistent account of the current human footprint is needed in order to evaluate the impact on the environments as well as risk for infrastructure. Identification of roads and settlements with satellite data is challenging due to the size of single features and low density of clusters. Spatial resolution and spectral characteristics of satellite data are the main issues regarding their separation. The Copernicus Sentinel-1 and -2 missions recently provided good spatial coverage and at the same time comparably high pixel spacing starting with 10 m for modes available across the entire Arctic. The purpose of this study was to assess the capabilities of both, Sentinel-1 C-band Synthetic Aperture Radar (SAR) and the Sentinel-2 multispectral information for Arctic focused mapping. Settings differ across the Arctic (historic settlements versus industrial, locations on bedrock versus tundra landscapes) and reference data are scarce and inconsistent. The type of features and data scarcity demand specific classification approaches. The machine learning approaches Gradient Boosting Machines (GBM) and deep learning (DL)-based semantic segmentation have been tested. Records for the Alaskan North Slope, Western Greenland, and Svalbard in addition to high-resolution satellite data have been used for validation and calibration. Deep learning is superior to GBM with respect to users accuracy. GBM therefore requires comprehensive postprocessing. SAR provides added value in case of GBM. VV is of benefit for road identification and HH for detection of buildings. Unfortunately, the Sentinel-1 acquisition strategy is varying across the Arctic. The majority is covered in VV+VH only. DL is of benefit for road and building detection but misses large proportions of other human-impacted areas, such as gravel pads which are typical for gas and oil fields. A combination of results from both GBM (Sentinel-1 and -2 combined) and DL (Sentinel-2; Sentinel-1 optional) is therefore suggested for circumpolar mapping.

Rapid developments in photon-counting and energy-discriminating detectors have the potential to provide an additional spectral dimension to conventional x-ray grayscale imaging. Reconstructed spectroscopic tomographic data can be used to distinguish individual materials by characteristic absorption peaks. The acquired energy-binned data, however, suffer from low signal-to-noise ratio, acquisition artifacts, and frequently angular undersampled conditions. New regularized iterative reconstruction methods have the potential to produce higher quality images and since energy channels are mutually correlated it can be advantageous to exploit this additional knowledge. In this paper, we propose a novel method which jointly reconstructs all energy channels while imposing a strong structural correlation. The core of the proposed algorithm is to employ a variational framework of parallel level sets to encourage joint smoothing directions. In particular, the method selects reference channels from which to propagate structure in an adaptive and stochastic way while preferring channels with a high data signal-to-noise ratio. The method is compared with current state-of-the-art multi-channel reconstruction techniques including channel-wise total variation and correlative total nuclear variation regularization. Realistic simulation experiments demonstrate the performance improvements achievable by using correlative regularization methods.