Search

With the popularity of geospatial applications, database updating is getting important due to the environmental changes over time. Imagery provides a lower cost and efficient way to update the database. Three dimensional objects can be measured by space intersection using conjugate image points and orientation parameters of cameras. However, precise orientation parameters of light amateur cameras are not always available due to their costliness and heaviness of precision GPS and IMU. To automatize data updating, the correspondence of object vector data and image may be built to improve the accuracy of direct georeferencing. This study contains four major parts, (1) back-projection of object vector data, (2) extraction of image feature lines, (3) object-image feature line matching, and (4) line-based orientation modeling. In order to construct the correspondence of features between an image and a building model, the building vector features were back-projected onto the image using the initial camera orientation from GPS and IMU. Image line features were extracted from the imagery. Afterwards, the matching procedure was done by assessing the similarity between the extracted image features and the back-projected ones. Then, the fourth part utilized line features in orientation modeling. The line-based orientation modeling was performed by the integration of line parametric equations into collinearity condition equations. The experiment data included images with 0.06 m resolution acquired by Canon EOS Mark 5D II camera on a Microdrones MD4-1000 UAV. Experimental results indicate that 2.1 pixel accuracy may be reached, which is equivalent to 0.12 m in the object space.

Current bathymetric models for the South China Sea (SCS) are largely based on predicted depths from gravity and sparse single-beam echo-sounder measurements. Such models lack high-resolution coastlines and shallow-water bottom features around atolls and islands. This study refines the gravity field of the SCS using sea surface heights from measurements of satellite altimeter Geosat/GM, ERS-1/GM, Jason-1/GM and the original Cryosat-2. A new one-minute gravity anomaly grid is determined. The modeled gravity anomalies show a 6-mgal RMS discrepancy with shipborne measurements in shallow waters. An altimeter-only bathymetric model is derived from the new gravity grid by the gravity-geological method that uses the latest global and regional models of the ocean depth and marine gravity as a priori knowledge. The new model outperforms current SCS bathymetric models and is accurate to 100 m, based on comparison with multi-beam depth measurements. Optical images from IKONOS-2, QuickBird-2, GeoEye-1, WorldView-1-2 and -3, are rectified and digitized to derive the zero (coastline) and 20-m depth contours (reef lines) around 44 atolls, which are integrated with the altimeter-only depths, giving significantly improved accuracies and spatial resolutions in modeled depths. The improvement percentages of coastlines by the satellite imagery range from 50% to 97% at 41 of the 44 atolls. We establish a webpage for free access to the optical and depth images, and the depth and gravity grids. We will continue to update satellite images, altimeter-derived gravity grids and bathymetric models over major atolls of the SCS.

Airborne LIDAR point clouds, which have considerable points on object surfaces, are essential to building modeling. In the last two decades, studies have developed different approaches to identify structure lines using two main approaches, data-driven and modeldriven. These studies have shown that automatic modeling processes depend on certain considerations, such as used thresholds, initial value, designed formulas, and predefined cues. Following the development of laser scanning systems, scanning rates have increased and can provide point clouds with higher point density. Therefore, this study proposes using topological elevation analysis (TEA) to detect structure lines instead of threshold-dependent concepts and predefined constraints. This analysis contains two parts: data pre-processing and structure line detection. To preserve the original elevation information, a pseudo-grid for generating digital surface models is produced during the first part. The highest point in each grid is set as the elevation value, and its original threedimensional position is preserved. In the second part, using TEA, the structure lines are identified based on the topology of local elevation changes in two directions. Because structure lines can contain certain geometric properties, their locations have small relieves in the radial direction and steep elevation changes in the circular direction. Following the proposed approach, TEA can be used to determine 3D line information without selecting thresholds. For validation, the TEA results are compared with those of the region growing approach. The results indicate that the proposed method can produce structure lines using dense point clouds.

This paper proposes a model-based method for texture mapping using close-range images and Lidar point clouds. Lidar point clouds are used to aid occlusion detection. For occluded areas, we compensate the occlusion by different view-angle images. Considering the authenticity of façade with repeated patterns under different illumination conditions, a selection of optimum pattern is suggested. In the selection, both geometric shape and texture are analyzed. The grey level co-occurrence matrix analysis is applied for the selection of the optimal façades texture to generate of photorealistic building models. Experimental results show that the proposed method provides high fidelity textures in the generation of photorealistic building models. It is demonstrated that the proposed method is also practical in the selection of the optimal texture.

This paper proposes a kernel-based strategy to analyze the distribution of local point clouds and detect 3D lines from dense airborne LIght Detection and Ranging (LIDAR) data. The proposed method, namely topological elevation connection (TEC) analysis, employs geometric criteria with an automatic procedure instead of threshold operation. The focused targets include flat roofs with parapets, step roofs, hipped roofs, and gable roofs. By comparing with reference data, the accuracy between the detected lines and reference data can reach within 0.2 m. Three-dimensional boundaries and structure lines are essential for building modeling and data registration. Many related works have focused on the collection of co-planar LIDAR points to calculate intersection lines with optimized thresholds. In addition, they did not focus on parapets because of the limitations of developed assumptions and the criteria used. It is crucial to analyze the local point distribution for the automation improvement of detection proces...

During the past decade, Taiwan has experienced an unusual and fast growing in the industry of mapping, remote sensing, spatial information and related markets. A successful space program and dozens of advanced airborne and ground-based remote sensing instruments as well as mobile mapping systems have been implemented and put into operation to support the vast demands of geospatial data acquisition. Moreover, in addition to the government agencies and research institutes, there are also tens of companies in the private sector providing geo-spatial data and services. However, the fast developing industry is also posing a great challenge to the education sector in Taiwan, especially the higher education for geo-spatial information. Facing this fast developing industry, the demands of skilled professionals and new technologies in order to address diversified needs are indubitably high. Consequently, while delighting in the expanding and prospering benefitted from the fast growing industry, how to fulfill these demands has become a challenge for the remote sensing and spatial information disciplines in the higher education institutes in Taiwan. This paper provides a brief insight into the status of the remote sensing and spatial information industry in Taiwan as well as the challenges of the education and technology transfer to support the increasing demands and to ensure the continuous development of the industry. In addition to the report of the current status of the remote sensing and spatial information related courses and programs in the colleges and universities, current and potential threatening issues and possible resolutions are also discussed in different points of view.

Edges and corners are two major image features in the modelling of man-made objects; an edge provides strong geometric orientation and corners possess good localisation. Feature detection is the basis of image processing for numerous applications such as image registration and object modelling. Completeness and localisation are the two major considerations for these applications; however, illumination, reflectance and shadows may interfere with image grey values to produce various gradients along an edge. Thus, threshold selection is an important step in obtaining suitable features in target-dependent methods as improper selection might cause information loss and broken edges. Instead of threshold selection, this study therefore proposes a feature extraction method using topological gradient connection (TGC) analysis involving three steps: grey value refinement, gradient computation and topological connection analysis. The first step uses a Gaussian filter to smooth the grey value image. The second step computes directional gradients to identify ridge pixels and collect feature candidates. The third step analyses adjacent candidates based on the criterion of topological connection. This three-step tracing procedure combines these connected candidates into a single object. The proposed scheme employs different images derived from various sensors and compares them with the Canny operator (using manually selected thresholds) and manually plotted corners for detection ability assessment. Experimental results indicate that the automatic results are more complete for subtle feature lines than the Canny edges. In addition, the proposed method provides higher flexibility in selecting suitable feature layers for different applications. Resume Les bordures et les coins sont, dans les images, des objets tres importants pour la modelisation d'objets artificiels; une bordure fournit une orientation geometrique robuste et les coins ont une bonne localisation. La detection de formes est la base du traitement d'images pour nombre d'applications, comme le recalage d'images ou la modelisation d'objets. La completude et la localisation sont deux criteres essentiels pour de telles applications; toutefois, l'illumination, la reflectance et les ombres peuvent influencer les niveaux de gris de l'image et produire des gradients variables le long d'une bordure. La selection des seuils est donc une etape importante pour l'obtention de formes correctes avec des methodes dependantes de l'objet, car une selection inappropriee peut causer la perte d'informations et des cassures dans les bordures. Plutot que la selection des seuils, cette etude propose ainsi une methode d'extraction de formes basee sur une analyse de la connexion du gradient topologique (TGC), en trois etapes: rehaussement des niveaux de gris, calcul des gradients, et analyse des connexions topologiques. La premiere etape utilise un filtre gaussien pour lisser l'image de niveaux de gris. La deuxieme etape calcule des gradients directionnels pour identifier les pixels de bordures et collecter les formes candidates. La troisieme etape identifie les candidats adjacents selon un critere de connexion topologique. Cette procedure de reconstruction en trois etapes regroupe en un seul objet les candidats ainsi connectes. Le schema propose utilise des images issues de differents capteurs et compare les resultats avec l'operateur de Canny (utilisant des seuils definis manuellement) et avec des coins pointes a la main, pour evaluer la qualite de la detection. Des resultats experimentaux montrent que les resultats de la methode automatique donnent des resultats plus complets que les bordures de Canny pour des formes lineaires fines. En outre, la methode proposee permet une plus grande flexibilite dans la selection de couches d'information de forme pour differentes applications. Zusammenfassung Kanten und Eckpunkte sind zwei bedeutende Bildmerkmale zur Modellierung kunstlicher Objekte: eine Kante liefert wichtige Information zur geometrischen Orientierung und Eckpunkte sind genau zu lokalisieren. Die Detektion von Merkmalen ist von zentraler Bedeutung fur die Bildverarbeitung, fur die Bildregistrierung und zur Objektmodellierung. Vollstandigkeit und Lokalisierung spielen dabei die wichtigste Rollen, allerdings konnen Beleuchtung, Reflexion und Schatten die Grauwerte eines Bildes beeinflussen und z.B. verschiedene Gradienten entlang einer Kante produzieren. Die Auswahl von Schwellwerten ist ein entscheidender Schritt, um geeignete Merkmale zu erhalten. Eine ungunstige Wahl eines Schwellwerts kann Informationsverlust oder unterbrochene Kanten bedeuten. Statt der Auswahl eines Schwellwerts wird eine Merkmalsextraktionsmethode vorgeschlagen, die eine Analyse der topologischen Verbindung von Gradienten (TGC) nutzt. Es werden hierzu drei Schritte bearbeitet: Grauwertverfeinerung, Berechnung der Gradienten und Analyse der topologischen Beziehungen. Mit einem Gauss Filter wird das Grauwertbild geglattet. Im zweiten Schritt werden Richtungsgradienten berechnet, um Kantenpixel zu identifizieren und geeignete Kandidaten auszuwahlen. Im dritten Schritt werden benachbarte Kandidaten auf der Basis der topologischen Verbindung analysiert. Durch die drei Schritte werde nmiteinander verbundene Kandidaten in ein Einzelobjekt uberfuhrt. Das vorgeschlagene Schema wird auf verschiedene Bilder von verschiedenen Sensoren angewandt. Die Ergebnisse werden mit dem Canny Operator (mit manuell gewahlten Schwellwerten) und mit manuell erfassten Eckpunkten verglichen, um die Qualitat der Merkmalserkennung zu prufen. Die experimentell erzielten Ergebnisse zeigen, dass das vorgeschlagene automatische Verfahren hohere Vollstandigkeit fur nicht sehr ausgepragte Merkmale erzielt, als die, mit dem Canny Operator extrahierten Kanten. Daruber hinaus bietet dieses Verfahren eine hohere Flexibilitat fur die Wahl geeigneter Merkmalsstufen in verschiedenen Anwendungen. Resumen Aristas y esquinas son las dos principales caracteristicas que modelan los objetos hechos por el hombre. Una arista proporciona una fuerte orientacion geometrica y las esquinas dotan de una buena localizacion. La deteccion de caracteristicas es la base de numerosas aplicaciones en proceso de imagenes como el registro de imagenes y el modelado de objetos. Completitud y localizacion son las dos mayores consideraciones para estas aplicaciones. Sin embargo la iluminacion, los reflejos y las sombras interfieren con el valor de gris de las imagenes produciendo variaciones de gradiente a lo largo de una arista. Para obtener caracteristicas apropiadas en metodos dependientes de dianas es importante la seleccion de un limite sobre el gradiente. La inadecuada seleccion de este puede causar perdida de informacion y la fractura de aristas. En lugar de la seleccion de un limite se propone un metodo de extraccion de caracteristicas usando el analisis de conexion de gradiente topologico (TGC) que tiene tres pasos: refinamiento de los valores de gris, calculo del gradiente y analisis de conexion topologica. En el primer paso se usa un filtro Gausiano para suavizar los valores de gris de la imagen. En el segundo paso se calculan los gradientes direccionales para identificar los pixeles en la cresta y seleccionar los candidatos a pixel de caracteristica. En el tercer paso se analizan los candidatos adyacentes usando criterios de conexion topologica. Este procedimiento de tres pasos combina los candidatos adyacentes en un unico objeto. Este esquema se ha aplicado en imagenes de distintos sensores y sus resultados se han comparado con los del operador de Canny (usando limites seleccionados manualmente) y manualmente se han trazado las esquinas para valorar la capacidad de deteccion. Los resultados empiricos indican que los resultados automaticos son mas completos para caracteristicas lineales sutiles que el operador de Canny. Ademas el metodo propuesto proporciona mayor flexibilidad en seleccionar caracteristicas apropiadas para aplicaciones diferentes.

Aerial imagery and LIDAR points are two important data sources for building reconstruction in a geospatial area. Aerial imagery implies building contours with planimetric features; LIDAR data explicitly represent building geometries using three-dimensional discrete point clouds. Data integration may take advantage of merits from two data sources in building reconstruction and change detection. However, heterogeneous data may contain a relative displacement because of different sensors and the capture time. To reduce this displacement, data registration should be an essential step. Therefore, this investigation proposes an edge-based approach to register these two data sets in three parts: (1) data preprocessing; (2) feature detection; and (3) data registration. The first step rasterizes laser point clouds into a pseudo-grid digital surface model (PDSM), which describes the relief with the original elevation information. The second step implements topological analyses to detect image edges and three-dimensional structure lines from the aerial image and PDSM. These detected features provide the initial positions of building shapes for registration. The third part registers these two data sets in Hough space to compensate for the displacement. Because each building may have prominent geometric structures, the proposed scheme transforms these two groups of edges, and estimates the correspondence by the Hough distribution. The following procedure then iteratively compares two groups of Hough patterns, which are from an aerial image and LIDAR data. This iterative procedure stops when the displacement is within a threshold. The test area is located in Taipei City, Taiwan. DMC system captured the aerial image with 18-cm spatial resolution. The LIDAR data were scanned with a 10-point density per square meter using the Leica ALS50 system. This study proposed a 50 cm spatial resolution of PDSM, which is slightly larger than the point spacing. The experiment selected two buildings to evaluate the performance of the proposed scheme. The manually edited building boundaries from the stereo aerial images are the reference data for validation. Comparisons indicated that the registration procedure could adjust the displacement within 50 cm, which relates to PDSM resolution. These preliminary results also demonstrated the possibility of providing locations for building reconstruction.

Image matching is a practical way to build up the relationship between point pairs. In remote sensing and photogrammetry applications, area-based matching algorithms are usually used under control environment, such as fixing exterior orientation parameters, to reach high matching precision. In order to improve the reliability of image matching, area-based matching algorithms are frequently performed with image pyramid and epipolar constraints. Thus, Central-Left-Right matching (CLR matching) is proposed to enhance the reliability. The method is designed to cope with unreliable matching when the objects are with surface discontinuity. The CLR matching puts a candidate point in the center, left and right windows. Thus, the elements of target window and search window may be corresponded to the same objects if the candidate point is located in a vertical edge. In addition, the directions of features in the image may be varied because of the object diversity. Thus, the CLR matching would be better to combine with multi-window operation in accordance with the feature direction. This paper analyzed the feature direction first followed by the multi-window matching. The major works in this investigation contain feature extraction, feature analysis, and image matching. The experiments compared the results of traditional image matching, CLR matching, and multi-window matching. The experimental results indicate that the proposed method improves the accuracy of image matching.

This study is motivated by the need for a micro-geographic information system (GIS) to represent and analyze 3D spatial data for the plotting of fire-fighting search and rescue routes within buildings. The GIS uses a 3D geometric network model (GNM) and the Dijkstra algorithm to consider smoke movement during different times of a building fire. Therefore, the route calculation algorithm can avoid routes through heavy smoke within buildings. In addition, when firefighters must search an area to find victims, the GNM and Ant Colony Optimization are applied to find the shortest path that passes through each room of the area. Finally, the GNM is implemented to perform a search and rescue route analysis from an actual underground station. The proposed method can not only provide the shortest safe route within a building but can also minimize the time required to search for potential victims.

A mobile mapping system (MMS) is developed based on a direct sensor orientation technique that integrates GPS/inertial measurement unit (IMU) and digital cameras. This system is capable of directly mapping geospatial information without the requirement of ground control. In this article, an MMS composed of two digital single-lens reflex cameras and an Applanix POS AV-510© is described. There are several systematic errors that need to be calibrated before achieving the goal of direct georeferencing. These are related to lens distortion, the lever-arm between the GPS and the camera's perspective center, and the boresight misalignment angles between the IMU and the camera frame. Experimental results show that the calibration field is the major error source. However, with the designed system, we can achieve a maximum three-dimensional positioning error of less than 18.27 and 7.74 cm at distances 20 m using a single camera with parallel imaging and dual-cameras with convergent imaging, respectively. The result...

Modern road systems have become complex networks with multiple layers, making three- dimensional (3D) road modeling an important task in the geoinformatic realm. Although traditional topographic maps contain explicit planimetric networks, they often lack sufficient elevation information to describe the vertical alignments in multi-layer road systems. In this investigation, we combine data from large-scale topographic maps and airborne light detection and ranging (LIDAR) data to reconstruct 3D road models. The proposed scheme includes two steps: planimetric networking and surface modeling. In the first part, road centerlines are determined then linked up and their topologies organized using the polylines extracted from large-scale topographic maps. In the second part, a filter is utilized for the extraction of road surface points from airborne LIDAR data. The three dimensional alignment of the profiles and cross-sections is then computed. Furthermore, to improve the realism of the road models, sur...

We integrate lidar point clouds and large-scale vector maps to perform building modeling. The proposed scheme comprises three major steps: (a) the preprocessing of lidar point clouds and vector maps, (b) roof analysis, and (c) building reconstruction. During the preprocessing stage, the building polygons are first obtained from the polylines, followed by the selections of lidar points in the building polygons. An irregular triangulated network is then built to represent the facets. The segmentation of planar facets for roof analysis is implemented by examining the patch size and the facet orientation. The interior 3D roof edges are then determined from the intersection of the roof planes. Finally, the building models are reconstructed through regularization. Two sample sites are tested for the purposes of validation. The experimental results indicate that the proposed scheme allows for high fidelity and accuracy, provided that the point cloud density is enough.

Extraction of a digital elevation model (DEM) from airborne lidar point clouds is an important task in the field of geoinformatics. In this paper, a new automated scheme that utilizes the so-called “climbingand- sliding” method to search for ground points from lidar point clouds for DEM generation is described. The new method has the capability of performing a local search while preserving the merits of a global treatment. This is accomplished by emulating the natural movements of climbing and sliding in order to search for ground points on a terrain surface model. To improve efficiency and accuracy, the scheme is implemented with a pseudo-grid data and includes a back selection step for densification. The test data include a dataset released from the ISPRS Working Group III/3 and one for a mountainous area located in southern Taiwan. The experimental results indicate that the proposed method is capable of producing a high fidelity terrain model.

This paper describes a method to generate true ortho-images from highresolution satellite images. Multi-view images are used to compensate for hidden areas. This work includes three steps: (1) the generation of traditional ortho-images using a terrain model, (2) the relief correction for buildings, and (3) the detection of, and compensation for, hidden areas. A rational function model (RFM) is used to determine the geometric correspondence between the image space and the object space. In the generation of traditional ortho-images, RFM parameters and a digital terrain model (DTM) are employed to rectify tilt displacements and relief displacements for the terrain. Then, using a digital building model (DBM), relief displacements for buildings are computed. The purpose of the detection of hidden areas is to avoid double mapping. Considering the satellite’s small field of view, an efficient method for the detection of hidden areas and building rectification is proposed. The test area covers the city of Hsinchu in northern Taiwan. The test images are obtained from the QuickBird satellite.

In this paper, we compare the geometrical performance between the rigorous sensor model (RSM) and rational function model (RFM) in the sensor modeling of FORMOSAT-2 satellite images. For the RSM, we provide a least squares collocation procedure to determine the precise orbits. As for the RFM, we analyze the model errors when a large amount of quasi-control points, which are derived from the satellite ephemeris and attitude data, are employed. The model errors with respect to the length of the image strip are also demonstrated. Experimental results show that the RFM is well behaved, indicating that its positioning errors is similar to that of the RSM.

In this paper we discuss a semiautomatic measuring system that can extract images of cracks in the surface of concrete from multitemporal images. An accuracy assessment and analysis of the relationship between expansion of the concrete and crack width are also included. The proposed scheme is comprised of four major components: (1) imaging geometry modeling for camera calibration; (2) initialization and skeletonization of cracks using the developed graphical user interface; (3) automatic crack tracing and width determination; and (4) multitemporal image processing and analysis. The experimental results indicate that the consistency of crack widths between the automatic extraction method and the manual measurement method could reach 0.05 mm or better.

In this paper, we describe an Adaptive Patch Projection scheme that can accelerate the orthorectification for satellite images without losing accuracy. The proposed scheme is comprised of two major components: (a) orbit modeling, and (b) image orthorectification. In orbit modeling, we provide a collocation procedure to determine the precise orbits. In image orthorectification, the area of interest is sequentially subdivided into four quadrate tiles until a specified threshold for terrain variations is met. The threshold of maximum terrain variation in a tile will be optimized according to the computational efficiency and the accuracy requirements. Once the ground tiles are determined, we perform adaptive patch projection to the corresponding image pixels. Test images from SPOTS Supermode and QuickBird satellites are included. The experimental results show that this algorithm can minimize the orthorectification computation time, while the modeling error is insignificant.

Due to its high spatial resolution, IKONOS imagery has the highest potential for large-scale topographic mapping of the resource satellites currently available. The orthorectification of IKONOS images thus becomes an important task. Since raw image data would, in general, not be provided by the satellite company, we have developed a method that performs the rectification of IKONOS GEO images. The proposed scheme includes the following steps: (1) the correction of relief displacement for ground control points (GCPs); (2) the performing of affine transformation between the ground coordinate system and the image coordinate system; (3) the performing of back projection for each pixel in the orthoimage; (4) the use of digital terrain model to calculate the relief displacement for the pixel in the previous step; and (5) image resampling. The experimental results indicate that the generated orthoimage may reach an accuracy of 2 m, when 6 GCPs are used.

tor has to estimate the location of hidden corners from the conThis paper presents a novel method for semi-automatically jugate images. This procedure is tedious and inefficient and constructing building models from photogrammetric 3D line has a limited accuracy, especially for connected buildings in segments of buildings, i.e., their roof edges. The method, which densely built-up areas. we call “Split-Merge-Shape” (SMS), can treat both complete Weidner (1997), Haala and Brenner (1998), and Brenner line segments as well as incomplete line segments due to image (2000) proposed the use of digital surface models (DSMS )t o reocclusions. The proposed method is comprised of five major construct 3D building models. The DSM data can be generated parts: (1) the creation of the Region of Interest (ROI) and pre- automatically using stereo-pairs, or can be obtained from airprocessing, (2) splitting the model by using the 3D line seg- borne laser scanning (Lohr, 1996). The problem is that precise ments to construct a combination of roof primitives, (3) merging buildingboundaries cannotbewell defineddue tothesegmenconnected roof primitives to complete the boundary of each tation of DSMS. Therefore, other complementary data, such as building, (4) shaping each building rooftop by connected ground plans of the building outlines, are necessary to assure coplanar analysis and coplanar fitting, and (5) quality assur- the reconstruction. This limits the practicability of the ance. The experimental results indicate that the proposed approach. method can soundly rebuild the topology from the 3D line In order to increase efficiency, Fischer et al. (1998) and segments and reconstruct building models with up to a 98

An original scheme to detect shoreline changes using multi-temporal satellite images and tidal measurements is presented here. First, the basic idea behind this investigation is to reconstruct a reference digital terrain model (DTM) for tideland areas from a set of SPOT satellite images sampled over a short period. Each image corresponds to a tidal measurement. Then, the shoreline, as interpreted from a historical satellite image, is compared with one traced from the reference DTM, according to the associated tidal elevations. Experimental results indicate that the area error of the test sand barriers ranges between 7.6% and 12.5%.

This paper describes an experiment for a proposed scheme using on­ board satellite data to perform three-dimensional positioning for a pair of SPOT stereostrips. Based on the sampling geometry of French SPOT satellite images, we first adjusted the on-board parameters by incorporating a small number of ground control points (GCPs). 1ben a procedure for three-dimensional modeling was per­ formed as follows. The central features of the proposed scheme are: (l) use of on­ board data to initialize orbit parameters and attitude data for the satellite; (2) cor­ rection of on-board parameters with low order polynomials; (3) use of reverse transformation to check the collinear relationship between object space and image space for the GCPs; (4) adjustment of attitude data for the satellite to minimize the error in terms of the GCP collinearity; and (5) space intersection. Experimental results indicate that, for a SPOT panchromatic stereostrip with a 0.6 base-to-height ratio and a 45 km by 280 km stereo model area, the root-mean-square-error for 3­ D positioning is less than 8 m in each direction when only 4 GCPs are used. Three-dimensional positioning for object points is an important task in the field of civil engineering when geometrical data processing is needed. Terrain data are often required in the planning, design, and construction stages of various transportation and hydraulic engineering projects (Epps and Corey 1990; Jones et al. 1990). Three-dimensional terrain information is also required for geometrical correction (Wiesel 1985) and terrain-dependent radiometric correction (Colby 1991) of remotely sensed images. Three-di­ mensional terrain data may be obtained from ground surveying with the ex­ pense of intensive labor, but a labor-saving approach is possible when pho­ tograrnmetric techniques are used. A higher efficiency may be anticipated by the use of a digital approach to automate the working procedure. Nowadays, because the spatial resolution of satellite images is significantly improved and stereoscopic sampling is available, the efficient extraction of 3-D terrain data becomes possible. Currently, French SPOT satellites have the highest tem­ poral and spatial resolutions among the available earth resource satellites. In addition, due to their side looking characteristics, SPOT satellites may pro­ vide cross-track stereo pairs. Thus, SPOT images have high potential for the extraction of 3-D terrain information. From the photograrnmetric point of view, collinearity condition equations are generally used in 3-D positioning for SPOT stereo pairs. Based on the equations, a bundle adjustment may be applied to model the satellite orien­ tation and the 3-D ground coordinates for a pair of conjugate image points (Konecny et al. 1987; Gugan and Dowman 1988). This approach may reach subpixel accuracy. However, a large number of ground control points (GCPs)

Fire departments already make use of geographic information systems (GIS) to reduce the response time in case of incidents. However, the complex internal structure of buildings on fire may make search and rescue difficult in emergency situations. Search is the act of firefighters' looking for something (e.g. the location of potential victims). Search is possibly one of themost dangerous activities on the fire ground. Sometimes the visibility is zero, and the environment is really hot. Because of limited operating time in the building, the key to a successful search is how quickly firefighters can search. This study is motivated by the need for a micro GIS to represent and analyze 3D spatial data in fire-fighting search operations. The method can minimize the searching time for potential victims. The developed method generates a 3D geometric network model from 2D building plans. To search a whole floor within the building, the genetic algorithmis used to find the optimal locations of rescue teams to deploy. In addition, when firefighters need to search an area in order to find the victims, the ant algorithm is applied to find the shortest path that passes through each room of the area. Finally, the model is implemented to perform fire-fighting search analysis froman underground station.

Assessment and inventory of natural hazards such as landslides are essential for effective watershed management and sustainable development. In Taiwan, a typhoon (tropical cyclone) or earthquake event can trigger hundreds to thousands of shallow landslides in mountainous areas with steep slopes and rapid streams. Therefore, how to improve the efficiency and accuracy of landslide mapping by means of GIS (geographic information system) and remote sensing techniques is an important research issue. This study proposes a novel, semiautomatic method for mapping and editing landslides at a watershed level. Data sources include airborne laser scanner (ALS) data and color/near infrared ortho-imagery: the ALS data provide topographic features such as elevation, slope, surface roughness, and object height, and the ortho-imagery furnishes the radiometric characteristic of land cover such as greenness index or NDVI (normalized difference vegetation index) for identifying bare grounds. Based on the derived topographic and radiometric parameters, the method first uses a global, automatic algorithm to interpret and delineate landslides. Then it uses a local region growing algorithm and a 3D eraser to edit and compile landslide maps. To explore the causes of mass movement, these landslide maps can also be registered with other geospatial data in a GIS for data visualization and analysis. Experimental results indicate that the method is highly efficient and accurate compared with results of human interpretation from the stereo pairs of aerial photographs. Because Taiwan experiences an average of four or five typhoons every year, this new, semiautomatic method is expected to provide a useful tool for watershed management.

This paper presents a systematic approach to utilize multi-temporal remote sensing images and spatial analysis for the detection, investigation, and long-term monitoring of landslide hazards in Taiwan. Rigorous orthorectification of satellite images are achieved by correction of sensor orbits and backward projections with ground control points of digital elevation models. Individual images are also radiometrically corrected according to sensor calibration factors. In addition, multi-temporal images are further normalized based on pseudo-invariant features identified from the images. Probable landslides are automatically detected with a change-detection procedure that combines NDVI filtering and Change-Vector Analysis. A spatial analysis system is also developed to further edit and analyze detected landslides and to produce landslide maps and other helpful outputs such as field-investigation forms and statistical reports. The developed landslide detection and monitoring system was applied to a study of large-scale landslide mapping and analysis in southern Taiwan and to the long-term monitoring of landslides in the watershed of Shimen Reservoir in northern Taiwan. Both application examples indicate that the proposed approach is viable. It can detect landslides effectively and with high accuracy. The data produced with the developed spatial analysis system are also helpful for hazard investigation, reconstruction, and mitigation.

We present here a new scheme that performs geometrical correction for airborne multi-spectral scanner imagery provided that an adequate reference image, i.e., an orthophoto, is available. Instead of the ray-tracing method, we determine the corresponding ground coordinates for each pixel of an input image by integrating feature point matching under Delaunay triangulated networks. The major components of the scheme include progressive generation of registration control points, and image registration. Comprehensive investigation for quantitative error analysis is also included. Experimental results indicate that the corrected images attain an accuracy of better than 09 pixels.

Satellite remote sensing technology and the science associated with evaluation of land use and land cover (LULC) in urban region makes use of the wide range images and algorithms. Yet previous processing with LULC methods is often time-consuming, laborious, and tedious making the outputs unavailable within the required time window. This paper presents a new image classification approach based on a novel neural computing technique that is applied to identify the LULC patterns in a fast growing urban region with the aid of 2.5-meter resolution SPOT-5 image products. Since some different classes of LULC may be linked with similar spectral characteristics, texture features and vegetation indexes are extracted and included during the classification process to enhance the discernability. The classifier is constructed based on the partial lanczos extreme learning machine (PL-ELM), which is a novel machine learning algorithm with fast learning speed and outstanding generalization performance. A validation procedure based on ground truth data and comparisons with some classic classifiers prove the credibility of the proposed PL-ELM classification approach in terms of the classification accuracy as well as the processing speed. It may be applied for "rapid change detection" in urban region for regular emergency response, regular planning, and land management in the future.

Image fusion is an important technique to integrate high resolution panchromatic image and low resolution multispectral image. The fused image enhances the capability for image interpretation. EMD (Empirical Mode Decomposition) is an effective method to decompose the linear or nonlinear signals into a set of IMFs (Intrinsic Mode Functions). The characteristics of EMD may apply to image fusion technique. The objective of this investigation is to establish a novel image fusion method using a 2-D EMD. The idea of image fusion based on EMD is to decompose the panchromatic and multispectral images into their IMFs. Then, we replace the high frequency IMF of multispectral image by high frequency IMF of panchromatic image. Finally, the image fusion is performed by reconstructing the mixed IMFs. The experimental results indicate that the proposed method may produce a fused image that preserves spatial and spectral information.

The development of digital aerial camera provides the possibility of acquiring highly overlapped aerial images with high spatial resolution. In addition to its high spatial resolution that improves the capability of image interpretation, the highly overlapped images provide favorable geometrical configuration with high redundancy. The high similarity of stereo images is, thus, beneficial to the reliable image matching. Hence, the 3-D point clouds from the image matching have the great potential in 3-D modeling. The topographic maps provide the distinct boundaries for building modeling. The strategies of building reconstruction with existing topographic maps may improve the quality, cost, and efficiency for building modeling. The objective of this investigation is to integrate highly overlapped aerial images and building boundaries from topographic map. The proposed semi-automatic method includes 3-D lines extraction and polyhedral model generation. In the beginning, the operator locates the initial lines by a graphical user interface. The initial lines are refined by stereo images. The precise 3-D lines are processed into a 3-D modeling by an inference engine. The experimental results indicate that the proposed method may reconstruct the 3-D building model effectively.

This paper presents systematic approaches to create photo-realistic three-dimensional (3D) digital city systems. In the created 3D digital city systems, prismatic and polyhedral building models are constructed from assorted remote sensing and spatial datasets, including topographic maps, aerial and satellite images, airborne and ground-based LIDAR point clouds. Close-ranged digital photographs and video sequences are used to generate facade texture images for photo-realistic texture mapping of the constructed building models. High performance visualization algorithms based on Level of Detail (LOD) processes are also implemented for real-time, interactive exploration and applications. A few applications of the constructed 3D digital cities to urban development evaluation, environmental simulation, hazard mitigation and crime scene reconstruction are presented to demonstrate the usability of the developed 3D digital city system.

The DTM derived by image matching delineates the canopy surface, rather than the ground surface, and is suitable for the radiometric correction for multispectral images. If a pair of multispectral images is available, the DTM may be automatically generated by image matching, the geometric/radiometric corrections may be performed in a selfsufficient environment. To fully utilize the complimentary information of multispectral images, a scheme for adaptively matching a pair of SPOT multispectral images is proposed. Before starting a matching procedure the raw images are transformed to the quasi-epipolar coordinate system. Based on the least squares matching, each of the three-band images is applied first for similarity assessment with its counterpart. Then the best match is selected among the three bands for each point of the stereopair. The best match is considered successful provided the correlation coefficient is larger than a threshold. In addition to matching the gray scale images, we also test the performance of matching first principal component (PC1) pairs. Experimental results indicate that the highest accuracy is reached when the adaptive scheme includes the PC1 matching.© (1994) COPYRIGHT SPIE--The International Society for Optical Engineering. Downloading of the abstract is permitted for personal use only.