Orthoimages can be used as control information to determine the exterior orientation of new aerial images. For that purpose, the image displacements on aerial images due to photographic tilt and topographic relief (including the height of topographic objects) should be exactly removed to generate "true" orthoimages. Thus the correct geometrical and radiometric information can be provided from them for control information. However it is always time-consuming and labour-intensive to capture the accurate heights of terrain area for removing the image displacements from topographic relief. Therefore, this paper will propose a methodology to generate the orthoimage patches from multiple aerial images without capturing accurate terrain heights. In the proposed method, a concept of floating plane, similar to the concept of floating mark, will be presented to produce several orthoimage candidate sets for a terrain area at some height range from multiple aerial images. Each orthoimage candidate set consists of several related orthoimage patches for a terrain area at some specific height. Subsequently, a procedure will be developed to decide an optimal orthoimage set for this terrain area among those orthoimage candidate sets. While the optimal orthoimage set is decided, the height information is also determined. After that, a unique orthoimage patch for this terrain area will be fused from this optimal ortlloimage set. The relevant assessment of accuracy is also discussed and conducted in this paper. The preliminary experiments prove the feasibility of the proposed method.

Monitoring soil moisture changes is of importance in many applications. Studies show that radar has great potential to provide useful near surface moisture mapping. By making use of the temporal correlation of the moisture changes, estimation and prediction might be achieved with satisfying accuracy. The temporal effects, when appropriately modeled, can be taken into account by means of the Kalman filter theory. The two equations in the Kalman filter, namely, process equation and measurement equation, are used simultaneously to estimate moisture content from radar measurements and to predict the moisture changes over time. The relationship of the moisture content and radar scattering coefficient is built into the measurement equation, while the dry-up (or wet-down) of soil moisture in multi-temporal states was modeled into the process equation. Validation of the presented approach is verified through numerical simulations of radar measurements over the bare soil surfaces. Simulation results indicate that the estimation and prediction of the multi-temporal moisture content can be improved when the dry-up or wet-down process is properly modeled.

Target identification and recognition of SAR images require good feature selection and enhancement. Due to the coherent process, it is difficult to discriminate the SAR target feature properties simply using the shape, shadow, tone, color and texture, to name a few. The scattering center is one of the important properties for extracting the SAR feature. In this paper, we modified an algorithm based a conjugate gradient (CG) optimum method originally proposed for spotlight mode SAR images, in order to enhance the Stripmap mode SAR targets. First, we introduced to the background of scatter center enhanced for Spotlight mode from early papers. Second, we presented the SAR received signal model as the basic of CG method. Following, most importance is to introduce how to construct a kernel of CG method. To validate the effectiveness and efficiency of the modified method, a series of RADARSAT SAR images at fine mode were tested with ground truth available overpass the image acquisition. In addition to, we import the fully polarization SAR data from NASNJPL AIRSAR acquired from south Taiwan when Sep. 27, 2000. We also compared the performance with MV (Minimum Variance) and MUSIC (Multiple Signal Classification) methods. Performance indices include target to clutter ratio, 3dB main lobe width and CPU time. From the logarithmic probability density distribution and column ordering plot of enhanced image and original image, it was demonstrated that modified method provides the best performance among the three methods besides CPU time.

Landslides are natural phenomena for the dynamic balance of earth surface. Due to the frequent occurrences of Typhoons and earthquake activities in Taiwan, mass movements are common threatens to our lives. Moreover, it is a common practice for the agencies of water reservoirs in Taiwan to make a reconnaissance of the landslides of the watershed every 5 to 10 years for the purpose of conservation. It is found that the application of aerial photo-interpretation technique for this purpose has been recognized as an effective approach since 1970s. However, an efficient and automatic interpretation scheme has never been established. Therefore, two issues are to be resolved for creating a useful and timely landslide database, i.e. the consistency of the sub-datasets and the completeness of the coverage. As the manual interpretation and automatic recognition are compared, the former is a practical and operational method, but the result it derived is largely dependent on the professional background of interpretation operator. In this paper, the interpretation knowledge is quantified into recognition criteria. Multi-source data, e.g. a Quickbird satellite image, DTM reduced from a LIDAR data, road and river vector data, are fused to construct the feature space for landslides analysis. Then, those features are used to recognize landslides by a multilayer perceptron (MLP) Neural Network Method. The extraction result is evaluated in comparison with the manual-interpretation result. The experiments indicate that the bare surface (color tone factor) is a significant interpretation key for landslides recognition. In this case study, the highest recognition accuracy for landslides and non-landslides are 98% and 88%, respectively. Therefore, the conducted method can assist landslides investigation efficiently and automatically.

Building model is one of the important spatial information of cyber city. The objective of this investigation is the shaping of the building models from LIDAR data. The major work includes building detection, boundary determination and roof reconstruction. First, we analyze the LIDAR data to detect the building regions. Then, the regular building boundaries are obtained by LIDAR contours. Finally, the roofs of the building are shaped from LIDAR point clouds. In the boundary determination, we use the contours to determine the building boundaries. It is assumed that building boundaries are orthogonal. The orthogonal and collinear constraints are applied in building boundary determination. In the roof reconstruction, a TIN-based region growing is applied to extract the roof planes. The coplanarity and adjacency between triangles are considered for TIN-based region growing. Then, a least squares regression is selected to determine the coplanarity function of planar segment. The proposed method has been tested with LIDAR data of Hsin-Chu Science-based Industrial Park with 1/1000 topomap in evaluation. Due to the data characteristic, we use comer-based and area-based accuracy assessment in the validation. Experimental results indicate that the horizontal error is better than 6% in terms of area, and the shaping error of the reconstructed models is better than 0.08 meter.

The positioning accuracy of the Global Positioning System (GPS) has been improved considerably during the past two decades. The main error sources such as ionospheric refraction, orbital accuracy, antenna phase center variation, signal multipath, and tropospheric delay have been reduced substantially, if not eliminated. In this study, seasonal effect on the GPS positioning accuracy is investigated. Taiwan is chosen for her unique geographic location and complex topography with abound water vapor in the air but spatially and seasonally dependent. The GPS data were collected from continuously operating reference stations by the Ministry of the Interior (MOl), Central Weather Bureau (CWB), and Industrial Technology Research Institute (ITRI) of Taiwan, and International GPS Service (IGS). The investigation of the relationship between variance of the vertical coordinate and change of climate is carried out by computing the GPS data collected in July and December 2003 to represent summer and winter, respectively. In addition, the comparison in results by using the Saastamoinen model and Hopfield model for correcting the influence of the atmospheric path delay is presented. It is found that the maximum difference in the monthly average ellipsoid height between the two months is about 20 mm. The corresponding daily maximum difference is 60 mm. Also, the ellipsoid height derived by using the Saastamoinen model is smaller than that by using the Hopfield mode1.

This study presents an application of digital video camera for object tracking and positioning.The main purpose of the study is to automatically trace and position a motionless object on the video images that are recorded on a moving vehicle. Because of the change of the camera's focal length and the movement of the vehicle, by observation, the appearance of a motionless object on the video images firstly will gradually change its shape and completely lose its shape after a period of time. The variation of the shape apparently will bring about the complexity for developing the automated algorithm to trace and position the object on the video image. In this study, we develop an object-based matching technique to implement the tracking task. The technique uses shape matrix algorithm (SMA) that has scale and rotation invariant characteristics to calculate the similarity of variant shapes between adjacent images. The technique consists of three steps:(1) the segmentation step that is used to separate object and background on the video image;(2) the matching step that uses SMA to match the variant shapes between adjacent images and obtain the relative orientation parameters of the shapes; and (3) the tracking step that employs both Target Define Ground Operator (TDGO) and relative orientation parameters to trace the feature point of the object. After the object is traced in every video frame, the photogrammetric collinearity condition is used to transform the object from the image coordinates to the ground coordinates. An experiment is performed to trace a motionless ship in the open sea. The result shows that the proposed method can successfully trace and position the ship even the ship had become entirely out of shape on the video images.

In various applications, areas of interest sometimes cover two or more satellite images, thus mosaicking for those images becomes important. Scene-based orbit adjustment could yield geometric discrepancies that could cause image discontinuity along the seam line during mosaicking. Thus, block adjustment is preferable. The objective of this investigation is to establish a simple yet effective block adjustment procedure for multi-orbit satellite images. The major works of the proposed scheme are:(1)adjustment of satellite's orbit accurately, (2)calculation of the error vectors for each tie point using digital terrain model and ray tracing technique, (3)refining the orbit using the Least Squares Filtering technique and (4)generation of the orthophotos. Test images are from SPOT 5 satellite. Experimental results indicate that, the relative accuracy between orthoimages is about 6 meter. It is also demonstrated that the mosaicked image is better seamed when tie points are employed in the adjustment.