Automatic matching of high‐resolution SAR images

Based on high‐resolution SAR data, in this paper, a novel automatic matching model is proposed. The model, which employs a coarse to fine strategy as a whole, consists of three steps. In the first step, edge features are extracted on different levels of pyramid images and an efficient Hausdorff distance‐based method is used to yield a coarse global feature match. Due to bi‐tree searching, the bottleneck of Hausdorff distance's matching is well resolved. Secondly, SSDA (Sequence Similarity Detection Algorithm) is employed to acquire tie‐points using a cross‐searching approach which treats features extracted from master and slave images equally. Finally, local‐adaptive splitting algorithm with MMSE (Minimum Mean Square Error) is used to achieve a fine matching; local‐adaptive splitting algorithm is the essential process to achieve sub‐pixel matching accuracy, which enhances the process's flexibility and robustness.

Airborne SAR images with high resolution are provided by the Institute of Electronics, CAS and used for experiments—the results of the experiments demonstrate that the model proposed in this paper is robust, with high accuracy (up to a fraction of a pixel), and can be successfully applied to automatic matching of high‐resolution SAR images.     

A Web‐based multi‐agent system for interpreting medical images

A difficult problem in medical image interpretation is that for every image type such as xray and every body organ such as heart, there exist specific solutions that do not allow for generalization. Just collecting all the specific solutions will not achieve the vision of a computerized physician. To address this problem, we develop an intelligent agent approach based on the concept of active fusion and agentoriented programming. The advantage of agentoriented programming is that it combines the benefits of objectoriented programming and expert system. Following this approach, we develop a Webbased multiagent system for interpreting medical images. The system is composed of two major types of intelligent agents: radiologist agents and patient representative agents. A radiologist agent decomposes the image interpretation task into smaller subtasks, uses multiple agents to solve the subtasks, and combines the solutions to the subtasks intelligently to solve the image interpretation problem. A patient representative agent takes questions from the user (usually a patient) through a Webbased interface, asks for multiple opinions from radiologist agents in interpreting a given set of images, and then integrates the opinions for the user. In addition, a patient representative agent can answer questions based on the information in a medical information database. To maximize the satisfaction that patients receive, the patient representative agents must be as informative and timely as communicating with a human. With an efficient pseudonatural language processing, a knowledge base in XML, and user communication through Microsoft Agent, the patient representative agents can answer questions effectively.     

Shadow detection in colour high‐resolution satellite images

Image shadow segmentation has become a major issue in satellite remote sensing because of the recent commercial availability of high‐resolution images. Detecting shadows is important for successfully carrying out applications such as change detection, land monitoring, object recognition, scene reconstruction, colour correction, etc. This paper presents a simple and effective procedure to segment shadow regions on high‐resolution colour satellite images. The method applies a region growing process on a specific band (namely, the c ₃ component of the c ₁ c ₂ c ₃ colour space). To gain in robustness and precision, the region expansion also imposes a restriction on the saturation and intensity values of the shadow pixels, as well as on their edge gradients. The proposed method has been successfully tested on QuickBird images acquired under different lighting conditions and covering both urban and rural areas.     

Mapping of salt‐affected soils using TM images

Landsat images supply significant information on the different aspects of the physical environment necessary to determine the distribution of salty soils. This work centres on the La Mancha area (Central Spain) which, due to its flat topography, exhibits ideal conditions to show surface and groundwater movement, as well as sediment distribution. Combination of these factors is essential for development. The main goal of this study is to derive information on the distribution of salt‐affected soils, as well as the area subject to flood risk. Multitemporal Landsat TM images selected from the 1990s have been used for this purpose. TM images are useful for delineation of mesozoic sediments rich in salts and/or areas with high humidity content. These images have been geo‐referenced to the UTM coordinate system and have been digitally enhanced. Principal components, NDVI and supervised classification were computed. Therefore, it is quite possible to predict with a high degree of probability the development of saline soils, but their precise classification is not always possible, owing to the fact that they are usually associated with other soil groups: calcisols, regosols, luvisols and kastanozems. Salt‐affected soils have been found to be associated with clayish mesozoic sediments, hollows, flood plain area and subsurface and groundwater movements. Complementary fieldwork is necessary in order to confirm digital values.     

Land use qualities identified in remotely‐sensed images

Land use can be defined as the intentional use of a specific piece of land resulting in patterns of ecological responses that are visible in land cover and landscape. The responses to land use often result in a heterogeneous combination of classes of land cover. Existing methods used in the classification of satellite imagery are limited in their capacity to handle categories consisting of heterogeneous or multiple land cover classes. Accordingly, a spatial relational post‐classification (SRPC) method has been developed which uses a spatial relational post‐classification of land cover classes based on the incorporation of information about identified land use qualities. This paper explains how this method works, and presents the results from a case study of the surroundings of Sötåsa village located in southern Sweden. Different land cover classes were aggregated semantically into two land use quality classes. In conclusion, it is argued that it is possible to make the semantic shift from reflectance to land use qualities using the developed method on satellite data, and that this provides considerable scope for the future analysis of land use.     

Fast over‐land atmospheric correction of visible and near‐infrared satellite images

A rapid atmospheric correction method is proposed to be used for visible and near‐infrared satellite sensor images over land. The method is based on a simplified use of a radiative transfer code (RTC), which is used only a priori, to generate Look‐Up‐Tables (LUTs) of the estimated surface reflectance. A typical scenario and ranges of values for the main atmospheric correction parameters are initially established. Each image pixel is treated as a slight deviation from the reference scenario defined by the vector of the typical values for the parameters. The assumption of the parameter's independence allows the use of one‐dimensional LUTs. The method is suitable for near real‐time processing or whenever a large number of data are to be handled rapidly. The operator intervention is minimal, and the computation time involved in the correction of a whole image is about 1000 times shorter than the full use of the base RTC. A test is performed with advanced very‐high‐resolution radiometer (AVHRR) visible and near‐infrared data, using the Second Simulation of the Satellite Signal in the Solar Spectrum (6S) RTC as the base code. The accuracy of the proposed method was compared with the standard use of the 6S RTC over the same dataset with resulting root mean square errors of 0.0114 and 0.0104 for AVHRR bands 1 and 2 for the estimated surface reflectance, respectively.     

Ontology‐based documentation of land degradation assessment from satellite images

In this paper, we introduce the idea of documenting operational chains for land degradation assessment using ontologies. We believe that this process will help end users better understand the application domain characteristics and evaluate the results of the assessment process. Since the application domain is wide, various operational chains for land degradation assessment and their associated documentation exist, according to different options. This parameterization process causes the development of different ontologies, which nonetheless are, to a certain extent, linked because of the common software components of the corresponding operational chains. We therefore propose a hierarchical structure of these ontologies; so that several requirements such as understanding of expert knowledge interconnections and of application domain variety, documentation and assimilation of new expert knowledge, and reusability of software components become feasible.     

Optimization in multi‐scale segmentation of high‐resolution satellite images for artificial feature recognition

Multi‐resolution segmentation, as one of the most popular approaches in object‐oriented image segmentation, has been greatly enabled by the advent of the commercial software, eCognition. However, the application of multi‐resolution segmentation still poses problems, especially in its operational aspects. This paper addresses the issue of optimization of the algorithm‐associated parameters in multi‐resolution segmentation. A framework starting with the definition of meaningful objects is proposed to find optimal segmentations for a given feature type. The proposed framework was tested to segment three exemplary artificial feature types (sports fields, roads, and residential buildings) in IKONOS multi‐spectral images, based on a sampling scheme of all the parameters required by the algorithm. Results show that the feature‐type‐oriented segmentation evaluation provides an insight to the decision‐making process in choosing appropriate parameters towards a high‐quality segmentation. By adopting these feature‐type‐based optimal parameters, multi‐resolution segmentation is able to produce objects of desired form to represent artificial features.     

A multi‐scale segmentation approach to filling gaps in Landsat ETM+ SLC‐off images

On 31 May 2003, the Landsat Enhanced Thematic Plus (ETM+) Scan Line Corrector (SLC) failed, causing the scanning pattern to exhibit wedge‐shaped scan‐to‐scan gaps. We developed a method that uses coincident spectral data to fill the image gaps. This method uses a multi‐scale segment model, derived from a previous Landsat SLC‐on image (image acquired prior to the SLC failure), to guide the spectral interpolation across the gaps in SLC‐off images (images acquired after the SLC failure). This paper describes the process used to generate the segment model, provides details of the gap‐fill algorithm used in deriving the segment‐based gap‐fill product, and presents the results of the gap‐fill process applied to grassland, cropland, and forest landscapes. Our results indicate this product will be useful for a wide variety of applications, including regional‐scale studies, general land cover mapping (e.g. forest, urban, and grass), crop‐specific mapping and monitoring, and visual assessments. Applications that need to be cautious when using pixels in the gap areas include any applications that require per‐pixel accuracy, such as urban characterization or impervious surface mapping, applications that use texture to characterize landscape features, and applications that require accurate measurements of small or narrow landscape features such as roads, farmsteads, and riparian areas.     

PAN‐sharpening of very high resolution multispectral images using genetic algorithms

A novel image fusion method is presented, suitable for sharpening of multispectral (MS) images by means of a panchromatic (PAN) observation. The method is based on redundant multiresolution analysis (MRA); the MS bands expanded to the finer scale of the PAN band are sharpened by adding the spatial details from the MRA representation of the PAN data. As a direct, unconditioned injection of PAN details gives unsatisfactory results, a new injection model is proposed that provides the optimum injection by maximizing a global quality index of the fused product. To this aim, a real‐valued genetic algorithm (GA) has been defined and tested on Quickbird data. The optimum GA injection is driven by an index function capable of measuring different types of possible distortions in the fused images. Fusion tests are carried out on spatially degraded data to objectively compare the proposed scheme to the most promising state‐of‐the‐art image fusion methods, and on full‐resolution image data to visually assess the performance of the proposed genetic image fusion method.     

Building‐based damage detection due to earthquake using the watershed segmentation of the post‐event aerial images

This paper presents an approach for detecting the damaged buildings due to earthquake using the watershed segmentation of the post‐event aerial images. The approach utilizes the relationship between the buildings and their cast shadows. It is based on an idea that if a building is damaged, it will not produce shadows. The cast shadows of the buildings are detected through an immersion‐based watershed segmentation. The boundaries of the buildings are available and stored in a GIS as vector polygons. The vector‐building boundaries are used to match the shadow casting edges of the buildings with their corresponding shadows and to perform assessments on a building‐specific manner. For each building, a final decision on the damage condition is taken, based on the assessments carried out for that building only. The approach was implemented in Golcuk, one of the urban areas most strongly hit by the 1999 Izmit, Turkey earthquake. To implement the approach, a system called the Building‐Based Earthquake Damage Assessment System was developed in MATLAB. Of the 284 buildings processed and analysed, 229 were correctly labelled as damaged and undamaged, providing an overall accuracy of 80.63%.     

Mapping recent deforestation in the Brazilian Amazon using simulated L‐band MAPSAR images

Brazilian Amazon Forest biomes are presently under intensive land cover conversion from natural vegetation to agriculture. Timely detection of recent deforestation through orbital remote sensing is a critical requirement for an operational land cover monitoring system in order to provide information to the regulatory systems and decision makers. Optical images present drawbacks for operation in the moist tropics and synthetic aperture radar (SAR) data are a real alternative. The feasibility of using multipolarized L‐band images simulating the Multi‐Application Purpose SAR (MAPSAR) satellite was examined for the detection of recent deforestation in the Tapajós region. The discrimination of recent deforestation from other land cover classes was evaluated through a quantitative analysis based on Jeffreys–Matusitas (JM) distances derived from training samples using amplitude values and supported by field survey. The investigation confirmed the possibility of the discrimination of recently deforested classes from other classes based on the L‐band images as proposed in the MAPSAR.     

Predicting slow‐drying fire weather index fuel moisture codes with NOAA‐AVHRR images in Canada's northern boreal forests

Fire danger predicted by the Canadian Fire Weather Index, a system based on point‐source weather records, is limited spatially. NOAA‐AVHRR images were used to model two slow‐drying fuel moisture codes, the duff moisture code and the drought code of the fire weather index, in boreal forests of a 250,000 km² portion of northern Alberta and the southern Northwest Territories, Canada. Temporal and spatial factors affecting both codes and spectral variables (normalized difference vegetation index, surface temperature, relative greenness, and the ratio between normalized difference vegetation index and surface temperature) were identified. Models were developed on a yearly and seasonal basis. They were strongest in spring, but had a tendency to saturate. Drought code was best modelled (R ² = 0.34–0.75) in the spring of 1995 when data were categorized spatially by broad forest cover types. These models showed improved spatial resolution by mapping drought code at the pixel level compared to broadly interpolated weather station‐based estimates. Limitations and possible improvements of the study are also discussed.     

Modelling the population density of China at the pixel level based on DMSP/OLS non‐radiance‐calibrated night‐time light images

The spatial distribution of population density is crucial for analysing the relationships among economic growth, environmental protection and resource use. In this study we simulated China's population density in 1998 at 1 km×1 km resolution by integrating DMSP/OLS non‐radiance‐calibrated night‐time images, SPOT/VGT 10‐day maximum NDVI composite, population census data and vector county boundaries. Population density, both inside and outside of light patches, was estimated for four types of counties, which were classified according to their light characteristics. The model for estimating population density inside the light patches was developed based on a significant correlation between light intensity and population, while the model for estimating population density outside of light patches was constructed by combining Coulomb's law with electric field superposition principle. Our method was simpler and less expensive than existing methods for spatializing population density. The results were consistent with other estimates but exhibited more spatial heterogeneity and richer information.     

A semi‐automated line tracing technique for monitoring ice margins in Antarctic images

The current shape and rate of change of the margins of the Antarctic ice sheet are poorly known. Since polar regions are inhospitable, remotely sensed images provide an invaluable data source for studying the ice margin. These images can be examined by a human expert or processed automatically by a computer. This paper describes the design and implementation of a semi‐automated technique to trace ice margins in Synthetic Aperture Radar images. The technique is based on a deformable contour model derived from the Kass et al. active contour model and Lobregt and Viergever's discrete dynamic contour. The contour is initialized by the user, to approximate the margin, and the deformation process causes the contour to accurately mould itself to the shape of the margin. The technique has been tested on a number of different images, with acceptable results in most cases.     

Predicting winter wheat condition,grain yield and protein content using multi‐temporal EnviSat‐ASAR and Landsat TM satellite images

Since optical and microwave sensors respond to very different target characteristics, their role in crop monitoring can be viewed as complementary. In particular, the all‐weather capability of Synthetic Aperture Radar (SAR) sensors can ensure that data gaps that often exist during monitoring with optical sensors are filled. There were three Landsat Thematic Mapper (TM) satellite images and three Envisat Advanced Synthetic Aperture Radar (ASAR) satellite images acquired from reviving stage to milking stage of winter wheat. These data were successfully used to monitor crop condition and forecast grain yield and protein content. Results from this study indicated that both multi‐temporal Envisat ASAR and Landsat TM imagery could provide accurate information about crop conditions. First, bivariate correlation results based on the linear regression of crop variables against backscatter suggested that the sensitivity of ASAR C‐HH backscatter image to crop or soil condition variation depends on growth stage and time of image acquisition. At the reviving stage, crop variables, such as biomass, Leaf Area Index (LAI) and plant water content (PWC), were significantly positively correlated with C‐HH backscatter (r = 0.65, 0.67 and 0.70, respectively), and soil water content at 5 cm, 10 cm and 20 cm depths were correlated significantly with C‐VV backscatter (r = 0.44, 0.49 and 0.46, respectively). At booting stage, only a significant and negative correlation was observed between biomass and C‐HH backscatter (r = ?0.44), and a saturation of the SAR signal to canopy LAI could explain the poor correlation between crop variables and C‐HH backscatter. Furthermore, C‐HH backscatter was correlated significantly with soil water content at booting and milking stage. Compared with ASAR backscatter data, the multi‐spectral Landsat TM images were more sensitive to crop variables. Secondly, a significant and negative correlation between grain yield and ASAR C‐HH & C‐VV backscatter at winter wheat booting stage was observed (r = ?0.73 and ?0.55, respectively) and a yield prediction model with a correlation coefficient of 0.91 was built based on the Normalized Difference Water Index (NDWI) data from Landsat TM on 17 April and ASAR C‐HH backscatter on 27 April. Finally, grain protein content was found to be correlated significantly with ASAR C‐HH backscatter at milking stage (r = ?0.61) and with Structure Insensitive Pigment Index (SIPI) data from Landsat TM at grain‐filling stage (r = 0.53), and a grain protein content prediction model with a correlation coefficient of 0.75 was built based on the C‐HH backscatter and SIPI data.     

Mapping forest successional stages following deforestation in Brazilian Amazonia using multi‐temporal Landsat images

Tropical forest successional stages have been mapped previously with multi‐temporal satellite sensor imagery. The precise identification and classification of such stages, however, has proved difficult. This Letter presents a new method for the classification of forest successional stages following deforestation in Brazilian Amazonia. Multi‐temporal Landsat Thematic Mapper (TM) and Enhanced Thematic Mapper Plus (ETM+) and derived fraction images and field data were used in a semi‐automatic classification approach. The results were encouraging and signal the application of the method for the entire Brazilian Amazonia.     

Sequential masking classification of multi‐temporal Landsat7 ETM+ images for field‐based crop mapping in Karacabey,Turkey

Three Landsat7 ETM+ images acquired in May, July and August during the 2000 crop growing season were used for field‐based mapping of summer crops in Karacabey, Turkey. First, the classification of each image date was performed on a standard per pixel basis. The results of per pixel classification were integrated with digital agricultural field boundaries and a crop type was determined for each field based on the modal class calculated within the field. The classification accuracy was computed by comparing the reference data, field‐by‐field, to each classified image. The individual crop accuracies were examined on each classified data and those crops whose accuracy exceeds a preset threshold level were determined. A sequential masking classification procedure was then performed using the three image dates, excluding after each classification the class properly classified. The final classified data were analysed on a field basis to assign each field a class label. An immediate update of the database was provided by directly entering the results of the analysis into the database. The sequential masking procedure for field‐based crop mapping improved the overall accuracies of the classifications of the July and August images alone by more than 10%.     

Spatial and temporal probabilities of obtaining cloud‐free Landsat images over the Brazilian tropical savanna

Remotely sensed data are the best and perhaps the only possible way for monitoring large‐scale, human‐induced land occupation and biosphere‐atmosphere processes in regions such as the Brazilian tropical savanna (Cerrado). Landsat imagery has been intensively employed for these studies because of their long‐term data coverage (>30 years), suitable spatial and temporal resolutions, and ability to discriminate different land‐use and land‐cover classes. However, cloud cover is the most obvious constraint for obtaining optical remote sensing data in tropical regions, and cloud cover analysis of remotely sensed data is a requisite step needed for any optical remote sensing studies. This study addresses the extent to which cloudiness can restrict the monitoring of the Brazilian Cerrado from Landsat‐like sensors. Percent cloud cover from more than 35 500 Landsat quick‐looks were estimated by the K‐means unsupervised classification technique. The data were examined by month, season, and El Niño Southern Oscillation event. Monthly observations of any part of the biome are highly unlikely during the wet season (October–March), but very possible during the dry season, especially in July and August. Research involving seasonality is feasible in some parts of the Cerrado at the temporal satellite sampling frequency of Landsat sensors. There are several limitations at the northern limit of the Cerrado, especially in the transitional area with the Amazon. During the 1997 El Niño event, the cloudiness over the Cerrado decreased to a measurable but small degree (5% less, on average). These results set the framework and limitations of future studies of land use/land cover and ecological dynamics using Landsat‐like satellite sensors.