Vol49_1_2006def 311 ANNALS OF GEOPHYSICS, VOL. 49, N. 1, February 2006 Key words hyperspectral remote sensing – urban land use – vegetation distribution – classification 1. Introduction The final aim of the work presented in this pa- per is to develop a methodology for the character- ization of an urban environment for environmen- tal monitoring. This topic is extremely interesting and urgent, given the huge amount of problems in urban areas related to environmental risks. These risks can be somehow quantified using remote sensed, and especially hyperspectral, data. Indeed, there are already some examples of urban hyperspectral data analysis. In Marino et al. (2000) the authors present urban material characterization, while in Xiao et al. (1999) veg- etation canopies and tree analysis in a so called «urban forest» is addressed. Land cover classifi- cation and sealed parts detection in urban areas are considered in Segl et al. (2000), Heiden et al. (2001), Roessner et al. (2001a). In all of these papers the need for hyperspectral data for cor- HySenS data exploitation for urban land cover analysis Fabio Dell’Acqua (1), Paolo Gamba (1), Vittorio Casella (2), Francesco Zucca (3), Jon Atli Benediktsson (4), Graeme Wilkinson (5), Andre Galli (6), Eva Savina Malinverni (7), Graeme Jones (8), Darrel Greenhill (8) and Lennart Ripke (8) (1) Dipartimento di Elettronica, Facoltà di Ingegneria, Università degli Studi di Pavia, Italy (2) Dipartimento di Ingegneria Edile e del Territorio (DIET), Facoltà di Ingegneria, Università degli Studi di Pavia, Italy (3) Dipartimento di Scienze della Terra, Facoltà di Scienze Matematiche Fisiche e Naturali, Università degli Studi di Pavia, Italy (4) Department of Electrical and Computer Engineering, University of Iceland, Reyjkavik, Iceland (5) General Faculty of Technology, Lincoln University, Lincoln, U.K. (6) Facoltà di Agraria, Università Politecnica delle Marche, Ancona, Italy (7) Facoltà di Ingegneria, Università Politecnica delle Marche, Ancona, Italy (8) Digital Imaging Research Centre, School of Computing and Information Systems, Kingston University, Surrey, U.K. Abstract This paper addresses the use of HySenS airborne hyperspectral data for environmental urban monitoring. It is known that hyperspectral data can help to characterize some of the relations between soil composition, vegeta- tion characteristics, and natural/artificial materials in urbanized areas. During the project we collected DAIS and ROSIS data over the urban test area of Pavia, Northern Italy, though due to a late delivery of ROSIS data only DAIS data was used in this work. Here we show results referring to an accurate characterization and classifica- tion of land cover/use, using different supervised approaches, exploiting spectral as well as spatial information. We demonstrate the possibility to extract from the hyperspectral data information which is very useful for envi- ronmental characterization of urban areas. Mailing address: Dr. Fabio Dell’Acqua, Dipartimento di Elettronica, Facoltà di Ingegneria, Università degli Studi di Pavia, Via Ferrata 1, 27100 Pavia, Italy; e-mail: fabio.dellacqua@unipv.it 312 Fabio Dell’Acqua et al. rectly accomplishing these tasks is clear, and the reasons may be summarized as follows: – Imaging spectrometry helps to character- ize urban sealing and soil composition, provid- ing in a unique way information useful to de- fine the risks of urban runoff and flooding. – Hyperspectral imaging may allow identi- fying and cataloguing potentially dangerous con- struction materials that have been used for cover- ings in the past (e.g., asbestos concrete). – Hyperspectral data acquired in the thermal infrared are useful in studies of urban heat effects, to establish the correlation between building ma- terial composition, heat flux and ground temper- ature. Our project aims at producing detailed land cover maps exploiting the potentials of hyper- spectral mapping, since it has been demonstrated (Roessner et al., 2001b) that significant improve- ments in classification of urban areas could be achieved using the spectral surface characteristics in the visible and near-infrared range, with a suit- able spectral resolution. More in detail, we want to propose precise and useful algorithms for: – Characterization and classification of land cover in urban areas using supervised approaches. – Correlation of the spectral characteristics of urban cover materials with their geometry and illumination conditions. – Definition of the vegetation status and corresponding risk maps for air pollution. – In summary, medium and high-resolution characterization of urban environment that in- cludes built structures, road networks, etc. Finally, it is worth noting that recent techni- cal studies show that more refined results can be obtained by combining hyperspectral measure- ments with those from different, complementary sensors or data sets (Madhok and Landgrebe, 1999) or by combining hyperspectral and radar measurements (Gamba and Houshmand, 2001; see also Nevatia et al., 1999). So, to make the study more effective, hyper- spectral data were collected over an urban area where a multi-sensor data set is already available (ERS, SIR-C/X-SAR, Landsat, IKONOS, EROS- A1, as well as multitemporal aerial data). An in- tegrated system has been designed, and is cur- rently under extensive testing, to manage this rather large data set (Costamagna et al., 2001). Data used in this work come from the EU funded HySenS project, proposed by the DLR (German Aerospace Agency) to make the ac- cess to research infrastructures easier. In this frame, free access to the data from DLR’s air- borne imaging spectrometers DAIS 7915 (Dig- ital Airborne Imaging Spectrometer) and RO- SIS (Reflective Optics System Imaging Spec- trometer) were offered to the European remote sensing community. Imaging spectrometer data sets acquired over test areas proposed by the in- dividual user groups were system corrected and calibrated to at-sensor radiance. DAIS and ROSIS have a spectral resolution capable of delineating the highly complex land uses, covers and patterns of urban environ- ments, uncovering environmental degradation, such as vegetation stress and soil/artificial ma- terials decomposition. However, environmental urban analysis is a new development in remote sensing (Ben-Dor et al., 1998), for which suit- able analysis tools are required. In the following we will show a more detailed description of the collected data set, the method- ology implemented in the project, and some pre- liminary results obtained with DAIS data, refer- ring to supervised classification approaches, ex- ploiting spectral as well as spatial information. 2. Data set The data used in this work correspond to the records of four flight lines over the urban area of Pavia, Northern Italy, acquired in the summer of 2002. The area was imaged by means of both DAIS and ROSIS, with different spatial and spectral resolution. The flight altitude was cho- sen as the lowest available for the airplane (about 1500 m), which resulted in the finest spa- tial resolution, i.e. 2.6 and 1.2 m for DAIS and ROSIS, respectively. The lines were chosen so that the higher resolution sensor (ROSIS) cov- ered the whole urban area. Therefore, the images from the DAIS sensor are partially overlapping, which will allow studying the effects of the di- rectional reflectance of urban materials on map- ping accuracy. Unfortunately, due to unprecedented prob- lems with the recorder of navigation parame- 313 HySenS data exploitation for urban land cover analysis ters, no ROSIS data was provided so far to our team and thus only DAIS data were analysed. During the flight campaign, many ancillary data were collected, and in particular a detailed inspection of the materials in portions of the ur- ban area was performed. Ground spectrometer and sun spectrometer data were collected as a support for the atmospheric correction proce- dure (relying on MODTRAN), which requires the availability of both ground and at-sensor spectra for a number of land cover cases. Only atmospherically corrected data were then used in this work. 3. Methodology Accurate georeferencing is required to make maps obtained from high spatial resolution data useful for comparison with other data sources and for integration into a GIS. For this reason one of the efforts of this work has been the choice of an algorithm to obtain such a result. The work was then devoted to information extraction from the HySenS imagery using dif- ferent approaches for urban and non-urban ar- eas, whose results were then combined. The main reason for processing the data in two parts is that the characteristics of scenes are very different. For the urban part, spatial struc- ture plays a very important role in discrimina- tion. For classification of the non-urban area, the spectral information is of prime importance. The main features were extracted from both data sets using two well established extraction methods, namely Discriminant Analysis Feature Extraction (DAFE) and Decision Boundary Fea- ture Extraction (DBFE) (Landgrebe, 2003). Classification algorithms were then applied to both the original data and the extracted features. For urban areas, the neuro-fuzzy spectral + +spatial approach outlined in Gamba and Del- l’Acqua (2003) was followed, and results com- pared with that of classical classification algo- rithms. For non-urban areas, instead, the data were assumed to be Gaussian distributed. A very lim- ited number of labeled training samples is avail- able, as the scene is mainly urban. Therefore, methods based on the use of enhanced statistics (Landgrebe, 2003) were applied. The enhanced statistics methods use some unlabeled samples to improve the estimates of the parameters used in the classification. The maximum likelihood classifier was used for classification. The methodology applied may be summa- rized in fig. 1, and corresponds to the structure of this paper, where the next section will be de- voted to geocoding issues, while Section 5 will provide some examples of classification results exploiting spatial as well as spectral character- istics of the urban environment. For the vegetation, NDVI was calculated over the entire image after averaging red and IR bands; the NDVI image was then processed to extract the Weighted Mean Patch Size (WMPS) index and the Lacunarity index which were then used to characterize the distribution of vegetation inside the considered urban area. 4. Geocoding To geocode the DAIS images, two non- parametric methods were experimented. A polynomial geometric correction was carried out using Ground Control Points (GCPs) select- ed manually from 1:2000 vector Topographic Maps and a polynomial cubic function. For the Fig. 1. The work flow of the research presented in this paper. 314 Fabio Dell’Acqua et al. flight line 2 we used 28 GCPs, distributed over the whole area obtaining the average RMS error of 1 pixel. The same procedure was followed for the flight line 3 with 36 GCPs resulting in an average RMS error of 0.91 pixel. To improve the accuracy another method was tested. The software PCI OrthoEngine 8.2 Satel- lite Project was used to carry out geometric cor- rections based on rational functions (see Dow- man and Dolloff, 2000), where the relation be- tween ground points and corresponding image points is defined as a ratio of polynomials. No system data is needed, but a DEM is necessary, as well as a minimum number of GCPs, related to the degree of the polynomial employed. We performed the geocoding of the image line 3 using 33 GCPs and a polynomial of the 2nd order, resulting in an average RMS error in X pixel coordinate of about 0.26 and Y pixel co- ordinate of about 0.27. The rectified images (line 2 and line 3) were resampled using nearest neighbour so as to re- tain the spectral scale values for subsequent multispectral classification. 5. Urban mapping After geocoding, urban land cover mapping and classification was considered. As a prelim- inary step, an analysis of the spectra of materi- als commonly found in the urban area was per- formed, using CNR spectral library. As a conse- quence, after proper radiometric and atmos- pheric calibration (based on MODTRAN, Berk et al., 1989), and taking into account the spatial ground dimension of the DAIS pixel, nine cov- er classes were considered, namely water, tree, asphalt, parking (asphalt with some concrete and soil), bitumen, bricks, meadow, bare soil, shadow. More refined classes are expected to be found after proper analysis of the results of this preliminary processing step of the data set. 5.1. Spectral and spatial land cover mapping To provide a quantitative evaluation of the classification results, a sample of the data was considered (fig. 2), and ground truth was pro- vided by available high resolution aerial and satellite images and with a ground survey. As already mentioned in the previous sec- tion, a complete analysis of the urban land cover cannot be obtained without considering both a spectral and a spatial characterization. Therefore, we applied either on the original data set or on the one reduced by means of the DAFE or BDFE approaches, the neurofuzzy classification chain introduced in Gamba and Dell’Acqua (2003). This is a two step approach, considering first a spectral analysis of the image with a neural ap- proach (spectral ARTMAP), and followed by a re-processing step taking into account the spatial patterns of the first classification map (spatial ARTMAP). The spectral analysis using DBFE processed data provided an overall accuracy of 94.3%, while the data coming from DAFE fea- ture extraction process provided a value of 96.7%. After the second step, these values were enhanced to 95.3% and 97%, respectively. See a visual comparison of spectral and spatial classi- fication results in fig. 3a,b and table I. Other supervised classification methods were tested on the same area. Results, in terms Fig. 2. A sample image of the test area used for classification, near the town center (north is to the bottom); this image was obtained by averaging sev- eral channels in the visible range of wavelengths. 315 HySenS data exploitation for urban land cover analysis Fig. 3a-d. Classification maps for a) the spectral ARTMAP and b) spatial ARTMAP classifier applied to the data after a DAFE analysis; classification maps for c) the ECHO and d) Maximum Likelihood classifier applied to the data after a DAFE analysis. a b c d Table I. Accuracy values for ARTMAP applied to the test area in fig. 2. Spectral ARTMAP Spatial ARTMAP Producer’s accuracy User’s accuracy Producer’s accuracy User’s accuracy Water 100% 100% 100% 100% Trees 93.6% 98.8% 94.2% 98.5% Asphalt 98.6% 96.9% 98.8% 96.6% Parking 96.9% 99.9% 90.9% 92.6% Bitumen 99.5% 99.9% 97.7% 99.7% Bricks 92% 87.7% 99.4% 99.9% Meadow 99.4% 82.7% 99.5% 82.6% Bare soil 86.9% 98.3% 87.6% 98.1% Shadow 90.9% 81.1% 87.3% 88.5% 316 Fabio Dell’Acqua et al. of classification accuracy, are shown in table II. The Spectral Angle Mapper (SAM) obtained poor results; this is probably due to the choice of the training set, optimized for statistical meth- ods but unsuitable for SAM which works best on pure spectra instead. Figure 3c,d and table III present the results for the Maximum Likelihood and ECHO classi- fiers applied after the DAFE feature extraction procedure. Table III reports the best of the two results, i.e. the one given by the Maximum Likelihood classifier. 5.2. Urban spatial structure investigation Together with accurate urban land cover maps, among the possible spatial characteriza- tions of an urban environment there is a growing need for quality of life indicators. In particular, the availability and distribution of green areas is a good indicator of both the air quality (which heavily affects the quality of life, especially in urban centers) and the possibility to have open- air entertainment areas. The approach taken in this work is to characterize the distribution of patches of vegetation in urban areas. As a first processing step, the pixels in our DAIS data set were classified into vegetative or non-vegetative pixels using the Normalized Dif- ference Vegetation Index (NDVI, Tucker and Sellers, 1986). The red bands 7 to 12 (0.63 to 0.70 mm) were averaged to make a red image, and the near infrared bands 16 to 21 (0.76 to 0.85 mm) were averaged to make a near infrared image. The NDVI image was then thresholded to create an image that contains patches of vegeta- tive pixels. These patches were then analyzed using two metrics: a weighted mean patch size metric and a Lacunarity metric. The Weighted Mean Patch Size (WMPS) in- dex (Li and Archer, 1997) calculates the patch area weighted by the number of patches in the scene possessing that area. An image was de- rived which shows the WMPS calculated over a local window within the image. A window is moved across the image in a raster scan fashion, and the WMPS is assigned to the central pixel in each window. The WMPS image thereby shows the spatial distribution of patches within the image. The Lacunarity index (Plotnick et al., 1993) shows the variability of size of vegetated patch- es, and is an indicator of «clumpiness». In a way similar to the WMPS, it has been applied to calculate a single value for each point in the scene. The Lacunarity index has an additional parameter known as the box size, which results in a Lacunarity image for each box size; results are shown in fig. 4a-d. These experiments should help provide new insights into landscape structure, which can be exploited in land use planning and in the cre- ation of heuristics for planning sustainable ur- ban development. Thresholding the Normalized Difference Vegetation Index (NDVI) pixels are classified into vegetative or non-vegetative pixels. The quality of the Lacunarity and Weighted Mean Patch Size (WMPS) metrics is determined by the robustness of this classification. The upper Table III. Accuracy values for Maximum Likeli- hood classifier applied to the test area in fig. 2. Producer accuracy User accuracy Water 100% 100% Trees 95.2% 98.3% Asphalt 96.2% 98.9% Parking 91.6% 82.2% Bitumen 98.3% 99.4% Bricks 99.8% 99.2% Meadow 98.4% 90.1% Bare soil 94.3% 97.7% Shadow 89.2% 82.7% Table II. Overall accuracy values for the test area in fig. 2. Original data DBFE DAFE Maximum Likelihood 95.6% 96.3% 97.6% Fisher 97% 96.7% 97% Minimum distances 73.6% 87% 96.6% ECHO 95.6% 96.3% 97.6% SAM with 0.05 rad 87.4% = = 317 HySenS data exploitation for urban land cover analysis and lower areas of the image (fig. 4a-d) contain the most vegetative pixels (displayed in white in fig. 4b). These regions also have the highest Lacunarity (fig. 4c) and WMPS (fig. 4d). The centre of the region is mostly urban, and has low values for both metrics. A large number of ecological metrics are available and it is important to establish which are the most useful. Early results from this work appear to suggest that for environmental- ly sustainable land use in sub-urban areas local WMPS should be maintained at values equiva- lent to 0.16 km2 and Lacunarity at values ex- ceeding 2.0 at box size 28 ×28 m and exceeding 1.4 at box size 204×204 m. These values are intended to ensure that a relatively low density of housing is maintained with a good clustering of local green areas, which would be support- ive of recreation and diversification of flora and fauna. 6. Conclusions and future work This work shows that it is possible to extract from hyperspectral data very important infor- Fig. 4a-d. a) The area of interest; b) NDVI image of the same area, thresholded at 0.5; c) Lacunarity image (window size 251, box size 9); d) WMPS image (window size 251). c d ba 318 Fabio Dell’Acqua et al. mation, useful for the environmental character- ization of urban areas. Even if in its preliminary stages, this research has shown many potential- ities for urban remote sensing. The co-registration of overlapping areas ac- quired by the DAIS in parallel flight lines having opposite directions opens the possibility to char- acterize urban materials considering also direc- tionality effects. This study will be helped by a number of precise ground reflectance measure- ments, made available thanks to CNR efforts. Acknowledgements The authors wish to thank the DLR for per- forming the campaign and providing the data within the framework of the HySenS project and the CNR for collecting the ground re- flectance data and the atmospheric calibration data and for their support. REFERENCES BEN-DOR, E., N. LEVIN and H. SAARONI (1998): Utilization of imaging spectrometry in urban areas: a case study over Tel-Aviv, Israel using the CASI sensor, in Pro- ceedings of the 1st EARSeL Workshop on Imaging Spectrometry, Zurich, Switzerland, 473-479. BERK, A., L.S. BERNSTEIN and D.C. ROBERTSON (1989): MODTRAN: a moderate resolution model for LOW- TRAN 7 (Phillips Lab., Hanscom AFB, MA), Tech. Rep. GL-TR-89-0122. COSTAMAGNA, E., P. GAMBA and V. CASELLA (2001): Envi- ronmental applications of the MURA DI PAVIA proj- ect, in Proceedings of the International Workshop on Geo-Spatial Knowledge Processing for Natural Re- source Management (CD-ROM). DOWMAN, I. and J.T. DOLLOFF (2000): An evaluation of ra- tional functions for photogrammetric restitution, in IAPRS, Amsterdam, vol. XXXIII. GAMBA, P. and F. DELL’ACQUA (2003): Improved multiband urban classification using a neuro-fuzzy classifier, Int. J.Remote Sensing, 24 (4), 827-834. GAMBA, P. and B. HOUSHMAND (2001): Integration of AVIRIS and IFSAR data for improved 3D urban profile reconstruction, Photogramm. Eng. Remote Sensing, 67 (8), 947-956. HEIDEN, U., S. ROESSNER, K. SEG and H. KAUFMANN (2001): Analysis of spectral signatures of urban surfaces for area-wide identification using hyperspectral HyMap data, in Proceedings of the IEEE/ISPRS Joint Work- shop on Data Fusion and Urban Remote Sensing over Urban Areas, Rome, Italy, 173-176. LANDGREBE, D.A. (2003): Signal Theory Methods in Multi- spectral Remote Sensing (John Wiley and Sons, Hobo- ken, New Jersey). LI, B.-L. and S. ARCHER (1997): Weighted mean patch size: a robust index for quantifying landscape structure, Ecol. Modelling, 102, 353-361. MADHOK, V. and D. LANDGREBE (1999): Supplementing hy- perspectral data with digital elevation, in Proceedings of the International Geoscience and Remote Sensing Symposium (IGARSS ‘99), Hamburg, Germany, June 28-July 2, 1999, IEE 1999 International, 1, 59-61, doi: 10.1109/IGARSS.1999.773400. MARINO, C.M., C. PANIGADA, L. BUSETTO, A. GALLI and M. BOSCHETTI (2000): Environmental applications of air- borne hyperspectral remote sensing: asbestos concrete sheeting identification and mapping, in Proceedings of the 14th International Conference and Workshops on Applied Geologic Remote Sensing, August 2000, Las Vegas, Nevada, U.S.A., CD-ROM. NEVATIA, R., A. HUERTAS and Z. KIM (1999): The MURI project for rapid feature extraction in urban areas, in IS- PRS: Automatic Extraction of GIS Objects from Digital Imagery, Munich, 3-14. PLOTNICK, R.E., R.H. GARDNER and R.V. O’NEILL (1993): Lacunarity indices as measures of landscape texture, Landscape Ecol., 8, 201-211. ROESSNER, S., K. SEGL, U. HEIDEN and H. KAUFMANN (2001a): Automated differentiation of urban surfaces based on airborne hyperspectral imagery, IEEE Trans. Geosci. Remote Sensing, 39 (7), 1525-1532. ROESSNER S., K. SEGL, U. HEIDEN and H. KAUFMANN (2001b): Analysis of spectral signatures of urban sur- faces for their identification using hyperspectral HyMap data, in Proceedings of the IEEE/ISPRS Joint Workshop on Data Fusion and Remote Sensing over Urban Areas, 8-9 November, Rome, Italy, 173-177. SEGL, K., S. ROESSNER and U. HEIDEN (2000): Differentia- tion of urban surfaces based on hyperspectral image data and a multi-technique approach, in Proceedings of the IEEE Geoscience and Remote Sensing Symposium, Honolulu, vol. 4, 1600-1602. TUCKER, C.J. and P.J. SELLERS (1986): Satellite remote sensing of primary production, Int. J. Remote Sensing, 7, 1395-1416. XIAO, Q., S.L. AUSTIN, E.G. MCPHERSON and P.J. PEPER (1999): Characterization of the structure and species composition of urban trees using high resolution AVIRIS data, in Proceedings of the AVIRIS Workshop, Pasadena, CA.