WSP02238 - Laserfiche WebLink

Home Browse Search

microns per picture element (1700 dpi) was adequate to retain all of the infomlation in the photographs (compare Figures 28 and 29); a scan resolution higherthan 15 microns produces too much image noise to be useful in photogrammetry. In addition, evaluations of various data compression engines during the remote-sensing initiativc determined that the best compression software was Gzip because it is lossless. frce-ware, and not copyrighted. At this time. the film library is being converted to digital format by the IT program and. since the beginning of the initiative. all image data collections have been delivered in digital format and have been duplicated for archival preservation. An ArclMS system has been implemented by the IT program which allows easy access to most of the collected digital data. 4.2.2. Most suitable image data for geomorphology Concerning future data collections, scientific review of image data collected during the remote-sensing initiative at different spatial resolutions showed Ihat 15-20-cm resolution was sufficient for geomorphic mapping of physical resources. We found that digital, orthorectified panchromatic imagery with 18-cm resolution can be acquired at about one-half the cost of color imagery. However, evaluations of different types of image data for mapping physical resources showed that CIR imagery discriminates surface materials better than panchromatic data (compare Figures 30 and 31) and allows more accurate digital classification of sand bars and debris tlows than does panchromatic imagery (Davis et aI., 2002b). In fact. CIR imagery (Figure 31) was found 10 be better than true-color imagery (Figure 29) for mapping sand bars and debris tlows, mainly because CIR imagery allows more complete removal of vegetation cover, which in turn allows more accurate determination of geologic surface textures (Davis et aI., 2002b). Surface texture was found to be a very important characteristic for distinguishing smooth sand bars and rough debris tlows (Davis et aI., 2002b); this was also found in other surficial-geologic studies (Shih and Schowengerdt. 1983; Anys et al.. 1994) 4.2.3. Airborne approaches for ground topography In March and in August-September of 2000, GCMRC collected L1DAR data with different collection parameters to evaluate the ability ofLlDAR to provide lopography at acceptable vertical accuracies on bare surfaces and on vegetated surfaces at the GCM RC sediment long-term monitoring sites. The March data were collected with the ALMS (Aeroscan Laser Mapping System) sensor at a spot spacing of 3.75 m and a spot diameter of 1.5 m. The August-September data were collected with the RAMS (Remote Airborne Mapping System) sensor at a l-m spot spacing and a 0.5-m spot diameter; Ihese data were acquired both in August and in September over the same four long-term monitoring siles using the same collection ~ parameters to evaluate the precision ofLlDAR elevation data. Photogrammetric data were derived from stereo image data (I A,OOO-scale photography) that were collected in September 2000 for one of the four study areas. Ground-survey elevation transects were performed at the sites during the L1DAR surveys and stakeout surveys were conducted at selected L1DAR point locations after the L1DAR surveys on different types of bare surfaces and within vegetated terrain. The L1DAR data sets and the photogrammetry data set were compared to the ground survey data to determine each data set's vertical accuracy. Various analyses were performed on these data over a two-year period (e.g., Davis et aI., 2002a; Mietz et al.. 2002), culminating in a final detailed evaluation of these data (Davis et al.. submitted). Our evaluations of high-resolution photogrammetry and of different rcsolution and replicate L1DAR data with respect to GCMRC monitoring requirements for sediment deposits provided the following conclusions; 24