This wall and car 2-in-1 battery charger kit charges the battery of Sony DCR-TRV738, DCR-TRV738E, DCR-TRV740, DCR-TRV740E, DCR-TRV75. Its slim, lightweight design makes the charger extremely convenient for travel use and it is one of the most compact battery chargers on the market. The home and car travel rapid charger charges battery in approximately 1 to 2 hours based on the battery capacity. Dimensions: 3.25 x 1.75 x 1.5 inch; Weight: 2.4oz. The charger is certified by CE & RoHS.
Part Numbers: Sony NP QM91D AC charger M200, Sony.NP.QM91D.AC.charger.M200
[ Report abuse or wrong photo | Share your Sony DCR-TRV738E photo ]
Sony DCR-TRV738E, size: 5.9 MB
Sony DCR-TRV738e Annexe 1
User reviews and opinions
No opinions have been provided. Be the first and add a new opinion/review.
3D MODEL GENERATION AND VISUALIZATION OF WARTBURG CASTLE
Helmut Mayera, Matthias Moschb, Juergen Peipec
Bundeswehr University Munich, D-85577 Neubiberg, Germany - Helmut.Mayer@unibw-muenchen.de Albert-Ludwigs-Universitt Freiburg, D-79106 Freiburg, Germany - email@example.com c Bundeswehr University Munich, D-85577 Neubiberg, Germany - firstname.lastname@example.org
KEY WORDS: Cultural Heritage, Multimedia, Internet, Calibration, Orientation, Modelling
ABSTRACT: In a joint project between Bundeswehr University Munich and Albert-Ludwigs-University Freiburg different methods to generate a 3D model of Wartburg Castle have been investigated. Several data acquisition devices were used, i.e. high resolution digital cameras and a low resolution camcorder. The virtual reconstruction of the historical building was accomplished by means of the low-cost (photogrammetric) software packages Canoma and PhotoModeler as well as by fully automatic computer vision methods for camera calibration, image orientation, and model generation. In this paper, the survey of Wartburg Castle is described and results of the modelling procedures are presented. 1. WARTBURG CASTLE results of the photogrammetric work concerning Wartburg Castle were presented in Mayer et al. (2003). Medieval castles are, in most cases, large and complex heritage sites showing a mixture of walls, gates, courtyards, towers, and buildings formed, extended, destroyed and reconstructed over centuries. Their survey is a relatively complicated and expensive task (e.g. Hanke & Oberschneider, 2002; Gonzo et al., 2004). In this paper, two image-based approaches for the documentation, 3D modelling and visualization of Wartburg Castle are presented. Subsequent to the acquisition of digital images (chapter 2), the photogrammetric object reconstruction was performed by the low-cost softwares Canoma and PhotoModeler (chapter 3) as well as by fully automatic procedures using computer vision techniques (chapter 4 and 5).
Founded in 1067, the Wartburg is a famous cultural heritage monument with regard to German history and an outstanding landmark of touristic interest. Wartburg Castle is located near Eisenach/Germany on top of a rocky hill and surrounded by the Thuringian Forest (Fig. 1 and 2). The terrain slopes steeply away on all sides of the castle complex. The inner courtyards are only accessible via a small drawbridge. The photogrammetric recording and 3D model generation is part of the EU funded research project Multimedia GeoInformation for E-Communities in Rural Areas with EcoTourism (Frech & Koch, 2003). Within the project an on-line tourist information system of the Thuringian Forest has been developed providing also visualization by means of, e.g., 2D tourist maps, aerial images, 3D models and 3D sceneries. First
Figure 1 and 2. Wartburg Castle
Rollei d30 metric Number of pixels Sensor format Lens (focal length) 2552 x mm x 7 mm
Canon EOS Dx mm x 15 mm 20 mm (SLR) ( = 32 mm for 35 mm camera) 7.4 MB uncompressed raw data per image
Sony DCR-TRV738E 720 x 576 3.6 mm - 54 mm ( = 48 mm - 720 mm fr 35 mm camera) 0.8 MB
10 mm - 30 mm ( = 40 mm - 120 mm for 35 mm camera) 6.4 MB uncompressed raw data per image
Table 1. Camera specifications
DATA ACQUISITION features such as camera calibration and bundle adjustment to achieve more accurate modelling results necessary in the courtyards and interior rooms of Wartburg Castle.
Digital images were acquired by means of three imaging systems, a 5 MP Rollei d30 metric5, a 6 MP Canon EOS D60 SLR camera, and a Sony camcorder (Tab. 1). The Rollei provides an optical zoom with two pre-calibrated focal lengths (15 mm and 30 mm). The camera was used to take some hundred photographs from the towers and windows of the castle, and from the ground. The Canon D60 with wide angle lens was suited to record several parts of the courtyards from a shorter distance. Additionally, the camcorder served to collect some overviews and short image sequences (movies) to be presented in the tourist information system, e.g., Luthers workroom and the famous banquet hall. Due to the location of the castle, images of the outer facades had to be taken from an ultra-light airplane. 3. OBJECT RECONSTRUCTION AND 3D MODELLING BY MEANS OF LOW-COST SOFTWARE The digital images were measured and object coordinates were determined within the low-cost software packages Canoma and PhotoModeler. The visualization tools of these programs were used to create a photorealistic model of the castle. Canoma (2004) was originally developed and sold by MetaCreations Corporation, meanwhile acquired by Adobe. Unfortunately the future status and availability of Canoma is not known. The software aims at modelling man-made structures, but not irregular shapes like trees or people. Canoma produces geometrically simple models for the Web and is not intended to replace detailed CAD models. Yet, it is easy to learn and to use, and one is able to get quick results, i.e. a photorealistic 3D model. Results can be posted on websites or exported into common 3D modelling and animation software via standard formats (e.g. DXF, VRML2). Also interactive walk-through animations and movies (recordable to Quicktime movie format) can be produced. Fig. 3 and 4 show images cut out of a virtual flight around Wartburg Castle. The well-known PhotoModeler software (Version 5.0; PhotoModeler, 2003; for camera calibration see also Wiggenhagen, 2002) was additionally applied to some regions of the castle. PhotoModeler offers advanced photogrammetric
Figure 3 and 4. Images extracted from a virtual flight around Wartburg Castle
FULLY AUTOMATIC ORIENTATION AND CALIBRATION
DETERMINATION OF PLANES AND THEIR EXTENT FOR THE INTERPRETATION OF BUILDING FACADES
The approach presented here is an extension of (Mayer et al., 2003). It rests on the scheme outlined in (Hao and Mayer, 2003). I.e., we match images in a pyramid and compute the uncalibrated orientation based on RANSAC (Fischler and Bolles, 1981) for pairs based on the fundamental matrix and for triplets based on the trifocal tensor (Hartley and Zisserman, 2003). The triplets are then linked via the direct linear transform (DLT; McGlone et al., 2004). Recent improvements comprise the substitution of RANSAC by an improved version termed geometric information criterion (GRIC; Torr, 1997) and the implementation of a multi image matching approach replacing the original pair-wise matching. The biggest change is the incorporation of the linear self-calibration approach of (Pollefeys et al., 2002, 2004) based on the dual absolute quadric making use of constraints on the parameters of a standard camera: It is assumed that the sheer is small, that the principal point is close to the center of the image and that the ratio of the focal length and the image size is in a range between one third and three. Everything together with a robust bundle adjustment taking into account radial distortion made it possible to automatically obtain for the Wartburg castle scene a calibration corresponding well to the known for the calibrated camera Rollei d30 metric5 and a back projection error in the range of 0.2 pixels. The latter is an improvement by a factor of two compared to the results reported in (Mayer et al., 2003). We wish to emphasize that the complete orientation and calibration is done totally automatically only based on the images and the assumption, that the images were taken by a perspective camera, possibly distorted by image deformations. There were no markers, ground control information, or approximate values used. The model needs to be rotated about three axes, shifted in three directions and scaled to comply with a real world coordinate system, but in spite of this, it is Euclidean, i.e., angles and distance ratios are correct. The bundle adjustment for nine images used for the results on the Wartburg castle scene given below was done with more than twenty five thousand observations and eight thousand unknowns from more than 2800 points, close to 1100 of them being 6-fold points and above.
We have recently obtained a research grant on building faade interpretation. Therefore, we have started to describe the Wartburg castle scene in terms of a faade model. This model basically consists of the description of faade parts by means of planes. The planes are detected from the relatively dense 3D points obtained by the orientation and calibration procedure described above by a GRIC based approach and then least squares matching is used to determine the actual extent of the planes. Before detecting planes, we orient the whole 3D scene in the vertical direction, facilitating a later detection of architectural elements such as windows and doors, which are usually vertically aligned. As we assume that we deal with architectural scenes there is always a larger number of vertical lines available, i.e., lines which intersect in the vertical vanishing point in the image. The user just has to tell the system if the vanishing point is in the x- or in the y-direction, and then the vanishing point for an image is determined by means of a RANSAC based approach for line intersections. We compute vertical vanishing points for one or more images and together with the known relative orientations of the images we determine an average vertical direction used for vertical alignment. To generate hypotheses for planes, we compute a large number of planes from randomly chosen triplets of points. Then, all points in the 3D scene are checked via the GRIC criterion, which takes into account also the precision of the 3D points obtained by the bundle adjustment, if they lie on a hypothesized plane. If the GRIC value lies below a threshold it is summed up, otherwise the threshold is summed. Finally, the hypotheses with the lowest sum is taken. As there can be more than one meaningful plane in our scene, we iterate the above procedure, generating additional planes which have a limited overlap in terms of common points (we use 30%) with already accepted plane hypotheses. The points defining the above planes give only a sparse representation of the faade. To determine their actual extent including the possibility of several topologically separated parts, but also to improve the precision of their parameters, we match the plane hypotheses in all images. We use the given orientation and plane parameters to project the image parts corresponding to the bounding box of the points defining the plane onto the plane. If all parameters were correct, the resulting images were perfectly aligned. Neglecting errors in the orientation of the cameras, differences between the projected images result from the unprecise plane parameters, but also from areas which do not lie on the plane. For faades this could be doors and windows lying behind the plane or drain pipes and columns lying in front of the plane. For more complex examples two parts of a faade residing on the same mathematical plane can also be separated by a part of the building. We treat both cases by means of robust least squares matching. The unknowns are the parameters of the plane as well as a bias and drift parameter for the brightness of every image. To separate areas actually residing on the plane from other areas, we use as threshold for the brightness difference to the average value of all images the double median
Figure 5. First courtyard of Wartburg Castle, partly used for the automatic measurement and modelling approach
value. To avoid a too strong bias from areas not on the plane on the parameters of the plane and particularly on the bias and drift parameter for every image, we start the adjustment with dilated areas around the points known to lie on the plane. Results of this approach on nine images of Wartburg castle with a size of 5 MP acquired with a Rollei d30 metric5 camera, the first image of which is given in Fig. 6, are shown in Figs. 7 and 8. As can be seen, we have extracted the seven major planes describing the two parts of the two faades, the roofs, and also the wall in front of the building. For the latter we also detected the holes. Please note, that due to limitations of computing resources the least squares matching was carried out in images of reduced resolution eliminating some details.
Our ultimate goal is to describe the faades in terms of its elements such as doors, windows, etc. To cope with the statistical nature of the problem, we want to use a statistical sampling approach based on reversible jump Markov chain Monte Carlo (RJMCMC) as proposed by (Dick et al., 2004) for faade interpretation. One interesting feature of this approach is, that one can simulate the model by sampling into the distribution without using image data, making it possible to check a model for its plausibility. One can also substitute parts not seen by a sensor, e.g., the back of a building, in a way consistent with the rest of the building.
Figure 6. First image of nine of Wartburg castle acquired with a Rollei d30 metric5 camera used for orientation, calibration, and faade plane estimation.
Figure 8. Frontal view, seen slightly from above of the same scene as in Fig. 7. On the lower side the green pyramids show the estimated positions of the cameras. For the wall in front of the building, the parts which have a pattern of holes can be distinguished from the solid parts.
Figure 7. Side view on the fully automatically generated 3D points as well as the planes computed by RANSAC and least squares matching. The green pyramids on the left side show the estimated positions of the cameras.
The use of 3D models of historical buildings has significantly increased over the last years. In this contribution the generation of a 3D model of Wartburg Castle as part of a tourist information system is described. Different image-based methods were applied, i.e. photogrammetric low-cost software and an automatic measurement and modelling approach. The latter allow fully automatic feature extraction, orientation, and 3D object reconstruction, even if the knowledge about the geometry of the image sequence is weak. Auto-calibration is possible in case that the interior orientation data have changed during the image recording (focusing, zoom) or are not available at all. It is advisable to combine several techniques to achieve virtual models from images as precise and reliable as required, especially in the case of a complex building like a medieval castle.
Hartley, R. and Zisserman, A., 2003. Multiple View Geometry in Computer Vision - Second Edition. Cambridge University Press, Cambridge, UK. Mayer, H., Mosch, M. and Peipe, J., 2003. Comparison of Photogrammetric and Computer Vision Techniques - 3D Reconstruction and Visualization of Wartburg Castle. In: The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol. (34) 5/C15, pp. 103-108. McGlone, C., Mikhail, E., Bethel, J. (2004): Manual of Photogrammetry Fifth Edition. American Society for Photogrammetry and Remote Sensing (ASPRS), Bethesda, USA. PhotoModeler, 2003. User Manual PhotoModeler Pro 5.0. Eos Systems Inc., 488 p. Pollefeys, M., Van Gool, L., Vergauwen, M., Verbiest, F., Cornelis, K. and Tops, J., 2004. Visual Modeling with a HandHeld Camera. International Journal of Computer Vision 59(3), pp. 207-232. Pollefeys, M., Verbiest, F. and Van Gool, L., 2002. Surviving Dominant Planes in Uncalibrated Structure and Motion Recovery. In: Seventh European Conference on Computer Vision, Vol. II, pp. 837-851. Wiggenhagen, M., 2002. Calibration of Digital Consumer Cameras for Photogrammetric Applications. In: The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol. (34) 3b, pp. 301-304. Torr, P., 1997. An Assessment of Information Criteria for Motion Model Selection. In: Computer Vision and Pattern Recognition, pp. 47-53.
REFERENCES Canoma Users.
Canoma, 2004. Resources for www.canoma.com (10 Sept 2004).
Dick, A., Torr, P. and Cipolla, R., 2004. Modelling and Interpretation of Architecture from Several Images. International Journal of Computer Vision 60(2), pp. 111-134. Fischler, M. and Bolles, R., 1981. Random Sample Consensus: A Paradigm for Model Fitting with Applications to Image Analysis and Automated Cartography. Communications of the ACM 24(6), pp. 381-395. Frech, I., Koch, B., 2003. Multimedia Geoinformation in Rural Areas with Eco-Tourism: The ReGeo-System. In: Information and Communication Technologies in Tourism 2003, Springer, Vienna/New York, pp. 421-429. Gonzo, L., El-Hakim, S., Picard, M., Girardi, S., Whiting, E., 2004. Photo-Realistic 3-D Reconstruction of Castles with Multiple-Sources Image-Based Techniques. In: The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol. (35) B5, pp.120-125. Hanke, K., Oberschneider, M., 2002. The Medieval Fortress Kufstein, Austria - An Example for the Restitution and Visualization of Cultural Heritage. In: The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol. (34) 5, pp. 530-533. Hao, X. and Mayer, H., 2003. Orientation and AutoCalibration of Image Triplets and Sequences. In: The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol. (34) 3/W8, pp. 73-78.
ACKNOWLEDGMENTS We want to thank Marc Pollefeys for making us available his MATLAB implementation of the linear self-calibration code based on the dual absolute quadric.
SM-308B DSC-S40 Cayenne GR-33 Pspbl PS-50Q91H SC-PM3 ADB341MP FX-9700GE L1246EL Pulse Smart 485 SR-38NMB KV3400DV2 Octave Samsung NX10 DVD-517H DP-3030 1200S NAS Funcub Dock 6000 ME-30 Minolta 2223 Acer M900 SP1604N-R NWZ-W252 T290I JS-P502 LM-4820C3L 23LB230B4U 3228C SPF AVR 10 Lide200 KV-28FQ86B Review 1200 FD S9600 KDL-52NX805 Chrysler 300C VT540 PRS-700 LG ZD30 HT-X725 Hasselblad 501C IC-2720H Minolta XG-1 VSX-609RDS Prosonic MDS-JE630 Azur 540A Simulink 7 KX-TGA721FX MDX-C8900R SPP-2040B CVP-205-CVP-205m-cvp-203 SDR-S26 Power Studio V-3310 25PT8303 Desktop PC HD950 Coolpix 5200 CR 20 Infocus IN82 SR9300 MP620 BD-HP21S RCR 127 USR5686E KW-AVX706 Zapco C2K CP-330 CPX1250 Blend Fostex MR-8 105VR 570LE-LB570be- MC-8483NL 27LB030B5 Software YP-S2QR 300 XC 2500W E-600 Sbchd1500U D16XD 3100CN Stylus 400 AC-VQ1050D P2015 Photosmart 370 Hd DVR KAC-646X 66331KF-N 91R 47LB2DE DGS-3427 STV1135 F2380M Ranger-2000 LX2060ST
manuel d'instructions, Guide de l'utilisateur | Manual de instrucciones, Instrucciones de uso | Bedienungsanleitung, Bedienungsanleitung | Manual de Instruções, guia do usuário | инструкция | návod na použitie, Užívateľská príručka, návod k použití | bruksanvisningen | instrukcja, podręcznik użytkownika | kullanım kılavuzu, Kullanım | kézikönyv, használati útmutató | manuale di istruzioni, istruzioni d'uso | handleiding, gebruikershandleiding
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101