Georeferenced Satellite Image Research Articles

Waypoint Generation in Satellite Images Based on a CNN for Outdoor UGV Navigation

Moving on paths or trails present in natural environments makes autonomous navigation of unmanned ground vehicles (UGV) simpler and safer. In this sense, aerial photographs provide a lot of information of wide areas that can be employed to detect paths for UGV usage. This paper proposes the extraction of paths from a geo-referenced satellite image centered at the current UGV position. Its pixels are individually classified as being part of a path or not using a convolutional neural network (CNN) which has been trained using synthetic data. Then, successive distant waypoints inside the detected paths are generated to achieve a given goal. This processing has been successfully tested on the Andabata mobile robot, which follows the list of waypoints in a reactive way based on a three-dimensional (3D) light detection and ranging (LiDAR) sensor.

Use of chirps satellite precipitation in the characterization of the franceses stream basin, Manaus - State of Amazonas - Brazil

The city of Manaus has numerous waterways in its urban perimeter, and, despite this, there is a shortage of measures aimed at protecting water resources that has generated occupations very close to the riverbeds. This work aims to characterize a watershed in an urban area of Manaus and map the areas subject to flooding around the Franceses Stream. The methodology applied involved obtaining topographic and thematic information such as morphology, geographical location, elevations, extent, area, level curves (water dividers) of the basin, obtained through georeferenced satellite images in a GIS environment.In addition, monthly satellite precipitation via the CHIRPS product was used for the calculations. The physical characteristics of the basin were determined using equations employed in hydrology. The results indicated that the Franceses Stream basin is not subject to major flooding, presenting the following characteristics: compactness coefficient Kc = 5.71; form factor Kf = 0.36; circularity index Ic = 0.03; drainage density Dd = 0.94km-1. It was also found that, despite the intense occupation, there is preserved vegetation within the basin. It was concluded that the precipitation estimated by CHIRPS is efficient because it has a larger coverage, and that knowing the characteristics of watersheds in urban environments is of great importance to subsidize intervention measures and planning policies that ensure the protection of water resources and minimize inconvenience to the population. It was also found that the use of GIS is an important tool for analyzing information in hydrological studies, and that it needs to become an increasingly applied tool in urban management, so that the knowledge obtained can be used to obtain better environmental and social conditions.

Detecting and Restoring Gaps Among Forest Patches: AN Achievable and Replicable Proposal to Boost the Landscape Connectivity

Abstract It is well known that one of the main impacts caused by land cover change is the process of forest fragmentation. Connecting the remaining fragments is always an exercise for technicians and academics, and even corridors of remaining vegetation usually present places of forest discontinuity, lacking connection strategies for the corridor to gain a real function. In this paper, we applied a model structured to identify priority locals to implement connectors. We used a GIS package and a digital, georeferenced satellite image. We complemented the project by surveying a database through a drone-based field survey. The integration of data revealed that for our studied area (Sorocaba, SP Brazil) 42.5% need assisted restoration. The mapping also revealed the existence of 25 forest fragments larger than 50 ha, arranged chiefly in one single corridor but disconnected from each other. Hence, through the application of the model, we could localize nine strategic locations in distances as short as possible among the fragments, feasible to implement a connector with the economy of resources and expect satisfactory performance in ecological terms. The database generated by the drone-based survey helped us to assert

Influence of environmental variability on the distribution and abundance of the pantropical spotted dolphin (Stenella attenuata) in the Mexican Central Pacific

The Mexican Central Pacific (MCP) has complex oceanographic dynamics, with a well-defined seasonal pattern and influence of interannual sea surface temperature anomalies. The present study investigated the relationship between the distribution/abundance of spotted dolphins and the environmental conditions in the MCP. Dolphin sighting data were collected during January 2010–December 2015; distance to coast, sea surface temperature, and chlorophyll a (Chla) were obtained for each sighting location through georeferenced maps and satellite images. A total of 746 spotted dolphin sightings were obtained with a survey effort of 31,695 km. Spotted dolphins in the MCP showed a predominantly coastal distribution regardless of environmental conditions. Significant differences were detected concerning regional seasonal variation, with higher mean sighting rate during the stratified periods and higher density/abundance estimates during the mixed periods. Spotted dolphins showed preference for warmer conditions, with higher mean sighting rate, abundance, and distance to coast during the neutral and warm interannual periods. During the coolest conditions (cold periods), mean sighting rate and abundance decreased and spotted dolphins were located closer to the coast, where Chla concentration increased. These results suggest that changes in the distribution/abundance of these dolphins could be associated with habitat quality related to tropical (El Niño/Southern Oscillation) and extratropical (The Blob) oceanographic phenomena, which could be modulating their foraging activities.

Census Mapping in India and Role of GIS: A Look Ahead at Census 2021

The 2021 Census of India for over 1.3 billion population deploying 3 million enumerators, has significant evidence value for 71 countries where census is scheduled during 2021. Census mapping plays a major role in accurate, complete and timely census. It delineates the exact and correct boundaries of all the administrative units. The Indian census has been using Geographic Information System (GIS) technologies over the last three censuses. In this study, we focus on the applications and methodologies being adopted for the census mapping in Census 2021 in India which is going to be the first digital Census of India. Five mobile apps have been developed for data collection and for map-related work. The 2021 Indian census utilises the latest census mapping techniques, namely standardisation of GIS spatial database design, geo-referencing of administrative units and latest mobile mapping application (Arc GIS Quick Capture) for field operations and built-up area digitisation work. We also discuss the various challenges and their solutions for census mapping in India, most prominently a high quality, updated, comprehensive and geo-referenced address registry for accurate data collection and mapping, and the use of geo-referenced high-resolution satellite images at village level for covering the gaps in rural boundary maps.

A New Method for Positional Accuracy Analysis in Georeferenced Satellite Images without Independent Ground Control Points

This paper introduces a new method to analyze the positional accuracy of georeferenced satellite images without the use of ground control points. Compared to the traditional method used to carry out this kind of analysis, our approach provides a semiautomatic way to obtain a larger number of control points that satisfy the requirements of current standards regarding the size of the set of sample points, the positional accuracy of such points, the distance between points, and the distribution of points in the sample. Our methodology exploits high quality orthoimages, such as those provided by the Aerial Orthography National Plan (PNOA)—developed by the Spanish National Geographic Institute—and has been tested on spatial data from Landsat 8. Our method works under the current international standard (ASPRS 2014) and exhibits similar performance than other well-known methods to analyze the positional accuracy of georeferenced images based on the use of independent ground control points. More specifically, the positional accuracy achieved for a Landsat 8 dataset evaluated by the traditional method is 5.22 ± 1.95 m, and when evaluated with the proposed method, it exhibits a typical accuracy of 5.76 ± 0.50 m. Our experimental results confirm that the method is equally effective and less expensive than other available methods to analyze the positional accuracy of satellite images.

Development of Bridges Maintenance Management System based on Geographic Information System Techniques (Case study: Al-Muthanna \ Iraq)

A Geographic Information System (GIS) is a computerized database management system for accumulating, storage, retrieval, analysis, and display spatial data. In general, GIS contains two broad categories of information, geo-referenced spatial data and attribute data. Geo-referenced spatial data define objects that have an orientation and relationship in two or three-dimensional space, while attribute data is qualitative data that can be counted for recording and analysis. The main aim of this research is to reveal the role of GIS technology in the enhancement of bridge maintenance management system components such as the output results, and make it more interpretable through dynamic colour coding and more sophisticated visualization techniques than the conventional tabular data format.
 To achieve the main objective of this research, two study areas have been chosen: the old constructionbridge (Al-Qadisiyah bridge) and the newly constructed bridge (Barboty bridge). Both of them are in Al-Muthanna city \ Iraq. The data collection process was achieved in two stages: the first stage is providing a georeferenced satellite image for each study area for the purpose of producing a two-dimensional map. The second stage includes the field surveying process by total station and level instruments.
 GIS have been used to create a comprehensive database (Geodatabase) for both study areas. Geostatistical analysis was carried out in which the settlement areas of both study areas were defined by producing a colour image. The statistical tables for these analyses showed that the highest decline in the elevation reached at Al-Qadisiyah bridge to 19 mm in the middle of the bridge which is coloured as a red areas. On the other hand, it was found that the highest decline in the elevation of the Barboty bridge is 16 mm in the last part of steel space which is also coloured as a red areas.

Application of Remote Sensing in Identifying and Mapping of Exposed Rock Massifs, Potential Quarry Sites in Yaounde-Cameroon

Since 2018, Cameroon opted for an aggressive mineral exploration in order to attain the status of an emerging country by 2035. Thus, any surface and sub-surface resources found important has to be valorised. This study therefore seeks to contribute towards the drive to an emerging nation initiative through the identification and mapping of exposed rock massifs in the city of Yaounde using Remote sensing techniques. The fieldwork constituted of locating the geographic coordinates of some rock outcrops with the GPS, projecting the samples on the a geo-referenced satellite image using GIS software (ArcGIS) and simulating the spectral behaviour of those spots to those of similar reflectance in order to conclude as the targets. This was then treated in GIS software for proper mapping to be done. The study revealed that Yaounde is actually endowed with enormous exposed rock massifs which can actually be valorised for construction works. Some of the exposed rocks are surrounded by vegetation especially at the outskirts of the city, others are actually engulfed within built-up areas with difficult accessibility while others could only be exposed on excavated areas for construction. Some are already under exploitation while others are not. Some of the rock outcrops were spotted of shallow extensions while others are small in surface areas but project deep down the soil. These results are of importance for use by quarry companies and the government of Cameroon through its Ministry of mines for onward allocation of exploitation permits.

LiDAR-based vehicle localization on the satellite image via a neural network

We present a novel method to localize the vehicle on an easily accessible geo-referenced satellite image based on LiDAR. We first design a neural network to extract and compare the spatial-discriminative feature maps of the satellite image patch and the LiDAR points, and obtain the probability of correspondence. Then based on the outputs of the network, a particle filter is used to obtain the probability distribution of the vehicle pose. This method can use LiDAR points and any type of odometry as input to localize the vehicle. The experimental results show that our model can generalize well on several datasets. Compared with other methods, ours is more robust in some challenging scenarios such as the occluded or shadowed area on the satellite image.

Development of Bridges Maintenance Management System based on Geographic Information System Techniques (Case study: AlMuthanna \ Iraq)

A Geographic Information System (GIS) is a computerized database management system for accumulating, storage, retrieval, analysis, and display spatial data. In general, GIS contains two broad categories of information, geo-referenced spatial data and attribute data. Geo-referenced spatial data define objects that have an orientation and relationship in two or three-dimensional space, while attribute data is qualitative data that can be counted for recording and analysis. The main aim of this research is to reveal the role of GIS technology in the enhancement of bridge maintenance management system components such as the output results, and make it more interpretable through dynamic colour coding and more sophisticated visualization techniques than the conventional tabular data format.
 To achieve the main objective of this research, two study areas have been chosen: the old construction bridge (Al-Qadisiyah bridge) and the newly constructed bridge (Barboty bridge). Both of them are in Al-Muthanna city \ Iraq. The data collection process was achieved in two stages: the first stage is providing a georeferenced satellite image for each study area for the purpose of producing a two-dimensional map. The second stage includes the field surveying process by total station and level instruments. GIS have been used to create a comprehensive database (Geodatabase) for both study areas. Geostatistical analysis was carried out in which the settlement areas of both study areas were defined by producing a colour image. The statistical tables for these analyses showed that the highest decline in the elevation reached at Al-Qadisiyah bridge to 19 mm in the middle of the bridge which is coloured as a red area. On the other hand, it was found that the highest decline in the elevation of the Barboty bridge is 16 mm in the last part of steel space which is also coloured as a red area.

Shoreline change identification and analysis along the coast of Guyana, South America

Guyana's coast is one of those in South America subjected to shoreline change. Understanding of spatio-temporal changes in shoreline position and the processes that drive the changes is very important in coastal protection, for increased shoreline surveillance and future planning. The aim of this study was to investigate the spatial and areal extent of this shoreline's morphological dynamics. It examined 49 years of the spatio-temporal behaviour of Guyana's shoreline by integrating and analysing the historical georeferenced satellite images in Google Earth repository, from 1969 to 2015 and onsite field mapping undertaken in 2017-2018. These were analysed using the Digital Shoreline Analysis System (DSAS) package of USGS in ESRI's© ArcGIS software. Rates of shoreline change were evaluated for the entire shoreline and these rates of change results varied at different sections of the entire coastline. Most of the Guyana coastal area continued to face shoreline erosion. The driving forces of the change were attributed to the natural forcing (sea level rise) and anthropogenic factors (e.g., sand mining and progressive depletion of mangrove forests). Anthropogenic interventions, such as the construction of breakwaters and mangrove rehabilitation and replanting were noticed to be responsible for some region-specific positive responses and adjustment to shoreline behaviour.

Predicting and Validating Malaria Vulnerability in Rural Kwara State, Nigeria: A Geospatial - Clinical Laboratory Approach

Background: Geographical Information System (GIS) establish relationship or link between vector borne diseases and associated environmental factors thereby providing explanation for the spatial distribution pattern, possible causes of diseases outbreak and implication on the community. We hypothesized, if GIS data could correlates with laboratory-based quantitative assessment of malaria prevalence pattern in rural Kwara State, Nigeria. This work focuses on utilization of GIS hardware, software to map, predict, communicate risk, in some selected geographic area and was substantiated with determination of malaria prevalence with RDTs and Giemsa Microscopy. Methods: Garmin GPS was used to capture the coordinates (Lat/Long) of 6 selected settlements and overlaid with a geo-referenced satellite image in the study area. A total of One hundred and thirty five (135) vulnerable individuals were diagnosed for Malaria with light Olympus microscope and rapid diagnostic kit. Data were entered and analyzed using R-Package for Statistical Computing and Graphics. Results: Apodu and Elemere had the highest malaria vulnerability index within 300m radius and the vulnerability index increases as one move away from the center of the settlement. The possible explanation for this high vulnerability could be the presence of pond/lake in Apodu. This is a good breeding site for mosquito couple with dense vegetation as one move from the centre of the settlements. Unlike Apodu, Gbugudu was at highest risk at 100 m buffer (60%) but the vulnerability index decreases as one move away from the settlement centre. The absence of thick vegetation and presence numerous cultivated farmlands on the eastern part could have been responsible for this observation. The RDT results revealed 37% malaria prevalance in all three communities combined while Light microscopy recorded 48% positivity rate. However, with Light microscopy, Apodu community had a higher infected to non-infected persons' ratio at 58.7% to 41.3% than Gbugudu and Elemere, both of which showed lower prevalence rates at 30.3% and 46.2% respectively. Conclusions: Findings of this study showed dense vegetation and ponds within Apodu, Elemere and Malete centres serves a good breeding site for mosquito coupled with dense vegetation was responsible for the high vulnerability index at these areas. Settlements within 1 km radius around KWASU campus recorded lower index possibly due to low vegetation. Results from this study indicate that the degree of malaria parasitaemia in the three settlement correlates directly with the remote sensing data. Funding Statement: Oluwasogo A Olalubi was a recipient of Tertiary Education (Tetfund) Institution Research Grant No:KWASUTETFUND/2016ABUJA/1/31. Declaration of Interests: No competing interests were disclosed. Ethics Approval Statement: All aspects of the study were approved by Kwara State University Research Committee and Ethical Review Board. Verbal and written Informed consent was obtained from all the community leaders/respondents in the three major settlement screened. They were assured of voluntary participation,confidentiality of their test results and responses and the opportunity to withdraw at any time withoutprejudice in line with the Helsinki Declaration was emphasized (World Medical Association, 2001).

Evaluation of Landscape Change Using Remote Sensing: A Case Study of Burdur Lake Basin

Georeferenced satellite images can be used for acquisition of topographic information, navigation and visualisation for various environmental studies. The objectives of this study were to determine land use and land cover changes and describe local landscape characteristics in the Burdur Lake Basin. Analyses were carried out using remotely sensed data. The topography of the land has been determined using GlobalDEM data of ASTER satellite with a resolution of 15 m. With this data, ArcMap 10.3 ArcHydro plugin was used to identify the basin of Burdur Lake and it was accepted as the boundary of the study. LT51790341987245 (02 August 1987) and LT51790342011254 (11 August 2011) coded Landsat TM satellite images have been used in the study. For the orthorectification, a digital elevation model (DEM) with an RMS accuracy of ca. 3.5 m, and for classifications 100 selected ground control points (GCPs) measured with differential GPS were used. Land uses identified with classified satellite images are analyzed in a spatial and proportional manner in a total study area of 3.361,56 km2. The changes in land cover are compared with each other and have been evaluated in terms of landscape change. Population growth, rising water consumption for agricultural and domestic purposes and building dams has led to water surfaces declining from 208,71 km2 to 155,42 km2. Also analyses show that the landscape had been dominated by a mixture of urban and cultivated land, and became more homogenous and aggregated in in 24 years.

Image matching applied to autonomous navigation of unmanned aerial vehicles

This paper brings results of an image matching approach applied to the estimation of position for autonomous navigation of unmanned aerial vehicles UAVs. The main idea is to replace the global positioning system GPS signal by matching the onboard video to a georeferenced satellite image. Image processing techniques and edge descriptors are considered in order to achieve robustness to different sensors and luminosity conditions. The UAV position is then calculated by finding the pixel in the georeferenced image providing a higher spatial correlation to the image captured at flight time. The evaluation of our vision system takes into account different kinds of terrains, such as in forests, roads and urban areas and its capacity to follow real routes specified in a flight simulator. The results obtained are promising and indicate that our methodology can be used in substitution for GPS on real flights.

A Unified Software Framework for Automatic Precise Georeferencing of Large Remote Sensing Image Archives

Though georeferencing satellite images is a solved industry problem long back, every satellite sensor data is unique to process due to its proprietary information and expected geometric accuracies usually demands use of reference data. Industry is moving towards unifying frameworks, sensor models where the users have to just plug-in details. The growing archive is inevitable and therefore (re) processing requirements have to handle huge image archives with state of art georeferencing standards automatically. Use of reference data is necessary to georeference the data sets to better accuracies beyond systematic corrections. This paper presents an approach for such tedious task of automatically reliably georeference huge image archives involving reference image/control points use in georeferencing Indian Remote Sensing satellite images.

Automatic geolocation of targets tracked by aerial imaging platforms using satellite imagery

Abstract. Tracking of targets from aerial platforms is an important activity in several applications, especially surveillance. Knowled ge of geolocation of these targets adds additional significant and useful information to the application. This paper determines the geolocation of a target being tracked from an aerial platform using the technique of image registration. Current approaches utilize a POS to determine the location of the aerial platform and then use the same for geolocation of the targets using the principle of photogrammetry. The constraints of cost and low-payload restrict the applicability of this approach using UAV platforms. This paper proposes a methodology for determining the geolocation of a target tracked from an aerial platform in a partially GPS devoid environment. The method utilises automatic feature based registration technique of a georeferenced satellite image with an ae rial image which is already stored in UAV's database to retrieve the geolocation of the target. Since it is easier to register subsequent aerial images due to similar viewing parameters, the subsequent overlapping images are registered together sequentially thus resulting in the registration of each of the images with georeferenced satellite image thus leading to geolocation of the target under interest. Using the proposed approach, the target can be tracked in all the frames in which it is visible. The proposed concept is verified experimentally and the results are found satisfactory. Using the proposed method, a user can obtain location of target of interest as well features on ground without requiring any POS on-board the aerial platform. The proposed approach has applications in surveillance for target tracking, target geolocation as well as in disaster management projects like search and rescue operations.

Automatic geolocation of targets tracked by aerial imaging platforms using satellite imagery

Abstract. Tracking of targets from aerial platforms is an important activity in several applications, especially surveillance. Knowled ge of geolocation of these targets adds additional significant and useful information to the application. This paper determines the geolocation of a target being tracked from an aerial platform using the technique of image registration. Current approaches utilize a POS to determine the location of the aerial platform and then use the same for geolocation of the targets using the principle of photogrammetry. The constraints of cost and low-payload restrict the applicability of this approach using UAV platforms. This paper proposes a methodology for determining the geolocation of a target tracked from an aerial platform in a partially GPS devoid environment. The method utilises automatic feature based registration technique of a georeferenced satellite image with an ae rial image which is already stored in UAV's database to retrieve the geolocation of the target. Since it is easier to register subsequent aerial images due to similar viewing parameters, the subsequent overlapping images are registered together sequentially thus resulting in the registration of each of the images with georeferenced satellite image thus leading to geolocation of the target under interest. Using the proposed approach, the target can be tracked in all the frames in which it is visible. The proposed concept is verified experimentally and the results are found satisfactory. Using the proposed method, a user can obtain location of target of interest as well features on ground without requiring any POS on-board the aerial platform. The proposed approach has applications in surveillance for target tracking, target geolocation as well as in disaster management projects like search and rescue operations.

DSM Accuracy Evaluation for the ISPRS Commission I Image Matching Benchmark

Abstract. To improve the quality of algorithms for automatic generation of Digital Surface Models (DSM) from optical stereo data in the remote sensing community, the Working Group 4 of Commission I: Geometric and Radiometric Modeling of Optical Airborne and Spaceborne Sensors provides on its website http://www2.isprs.org/commissions/comm1/wg4/benchmark-test.html a benchmark dataset for measuring and comparing the accuracy of dense stereo algorithms. The data provided consists of several optical spaceborne stereo images together with ground truth data produced by aerial laser scanning. In this paper we present our latest work on this benchmark, based upon previous work. As a first point, we noticed that providing the abovementioned test data as geo-referenced satellite images together with their corresponding RPC camera model seems too high a burden for being used widely by other researchers, as a considerable effort still has to be made to integrate the test datas camera model into the researchers local stereo reconstruction framework. To bypass this problem, we now also provide additional rectified input images, which enable stereo algorithms to work out of the box without the need for implementing special camera models. Care was taken to minimize the errors resulting from the rectification transformation and the involved image resampling. We further improved the robustness of the evaluation method against errors in the orientation of the satellite images (with respect to the LiDAR ground truth). To this end we implemented a point cloud alignment of the DSM and the LiDAR reference points using an Iterative Closest Point (ICP) algorithm and an estimation of the best fitting transformation. This way, we concentrate on the errors from the stereo reconstruction and make sure that the result is not biased by errors in the absolute orientation of the satellite images. The evaluation of the stereo algorithms is done by triangulating the resulting (filled) DSMs and computing for each LiDAR point the nearest Euclidean distance to the DSM surface. We implemented an adaptive triangulation method minimizing the second order derivative of the surface in a local neighborhood, which captures the real surface more accurate than a fixed triangulation. As a further advantage, using our point-to-surface evaluation, we are also able to evaluate non-uniformly sampled DSMs or triangulated 3D models in general. The latter is for example needed when evaluating building extraction and data reduction algorithms. As practical example we compare results from three different matching methods applied to the data available within the benchmark data sets. These results are analyzed using the above mentioned methodology and show advantages and disadvantages of the different methods, also depending on the land cover classes.

Orientation of churches by magnetic compasses?

SUMMARY In Christian religion the sunrise is of great symbolic importance. Therefore, many churches constructed in the Middle Ages point towards geographic East. Although ‘easting’ of churches actually refers to the alignment towards the azimuth of sunrise on the individual churches’ patron’s day, deviation of nave alignment from the geographic East direction is often assumed to be caused by the use of magnetic compasses. Therefore, the church alignment could provide information about historical magnetic declination. We investigate 124 churches in Lower Austria and 68 in northern Germany to clarify this question as well as the ‘easting’ hypothesis. Church orientations are determined from georeferenced satellite images. Metadata such as the construction year, possible reconstructions and the church patron are gathered to determine the date when current church direction was appointed, and to perform sunrise calculations. However, due to uncertainties of construction years and the declining importance of orientation tradition after the 15th century several churches are excluded from the study. Thus, 32 churches with reliable metadata remain for evaluation in each region. The analysis reveals a preferred alignment of naves towards geographic East in Lower Austria and northern Germany. The construction and alignment of churches was often affected by the pre-existing buildings and streets or topography and natural surroundings. Therefore, deviations from geographic East are more likely caused by town or landscape. The mean deviations from magnetic East in both regions are large compared to the mean deviations from geographic East and sunrise azimuths. Hence, the use of compasses cannot be confirmed. Despite a few churches indicating orientation according to their patron’s day, a general trend cannot be observed in the data reviewed.

Spatial and temporal changes in bird assemblages in forest fragments in an eastern Amazonian savannah

We investigated the effects of forest fragmentation on bird assemblages in an Amazonian savannah landscape with forest fragments that have been isolated for more than 100 years. The study was conducted in areas surrounding the village of Alter do Chão (2°31'S, 55°00'W), Santarém, Brazil. Bird surveys and measurements of tree density were undertaken in 25 areas, with 19 plots in forest fragments of different sizes and six in an area of continuous forest. Data on forest-fragment size, perimeter, and isolation were obtained from a georeferenced satellite image. Variation in number of bird species recorded per plot was not related to vegetation structure (tree density). The number of bird species recorded per plot increased significantly only with fragment area, but was not influenced by fragment shape or degree of isolation, even when considering species from the savannah matrix in the analysis. Fragments had fewer rare species. Multivariate ordination analyses (multiple dimensional scaling, [MDS]) indicated that bird species composition changed along a gradient from small to large forest fragments and continuous-forest areas. In the Amazonian savannah landscapes of Alter do Chão, the organization and composition of bird assemblages in forest fragments are affected by local long-term forest-fragmentation processes. Differences in the number of bird species recorded per plot and assemblage composition between forest fragments and continuous forest were not influenced by forest structure, suggesting that the observed patterns in species composition result from the effects of fragmentation per se rather than from preexisting differences in vegetation structure between sites. Nevertheless, despite their long history of isolation, the forest fragments still preserve a large proportion (on average 80%) of the avifauna found in continuous-forest areas. The fragments at Alter do Chão are surrounded by natural (rather than planted) grassland, with many trees in the savannah matrix and the landscape has vast areas covered by forest, which may have helped to ameliorate the influences of forest fragmentation.