SEGMENTATION QUALITY ASSESSMENT FOR VARYING SPATIAL RESOLUTIONS OF VERY HIGH RESOLUTION SATELLITE IMAGERY
T.Kavzoglu, H.Tonbul
Gebze Technical University, Engineering Faculty, Department of Geomatics Engineering,
41400, Kocaeli, Turkey
Abstract. Due to the complex nature of remotely sensed imagery, it is difficult to construct meaningful image objects by segmenting a landscape features in an image. Because many factors including parameter selection, band weights, spectral resolution, spatial resolution and textural information affect the quality of the segments to be produced, a comprehensive analysis is required to assure high quality image objects. In this study, the influence of the spatial resolution on segmentation quality was analysed using Worldview-2 satellite image at five different spatial resolutions (0.5, 2, 4, 8, 16 meters). The multiresolution segmentation algorithm, the most widely used method and available in eCognition software, was utilized for the segmentation processes in this study. The effect of spatial resolution on the segmentation quality was investigated on three specific land use/cover types namely, building, pasture and road by using quality measures of shape index, area fit index and quality rate. It has been observed that resampling the image from 0.5 to 2, 4, 8, 16 meters remarkably reduced the quality of the segmentation results. For instance, when increasing the spatial resolution from 8 to 16 meters, the quality rate decreased by about 77% for road class. The results of this study revealed that the use of 4 meters or higher resolutions (i.e. 0.5 and 2 meters) would produce acceptable results in terms of segmentation quality metrics. When the lower resolution is preferred, the quality of the segments decreases considerably, thus the created image objects become too coarse, indicating an increase in under-segmentation.
REFERENCES
1. Baatz, M., Schape, A., 2000, Multiresolution Segmentation: An Optimization Approach for High Quality MultiScale Image Segmentation. Strobl, J., Blaschke, T. and Griesbner, G. (Ed.), Angewandte Geographische Informations- Verarbeitung, XII, Wichmann Verlag, Karlsruhe, Germany, 12-23.
2. Cheng, J., Bo, Y., Zhu, Y., Ji, X., 2014. A novel method for assessing the segmentation quality of high-spatial resolution remote-sensing images. International Journal of Remote Sensing 35 (10), 3816–3839.
3. Clinton, N., Holt, A., Scarborough, J., Yan, L., Gong, P., 2010. Accuracy Assessment Measures for Object-based Image Segmentation Goodness. Photogrammetric Engineering & Remote Sensing 76 (3), 289–299.
4. Drăgut, L., Csillik, O., Eisank, C., Tiede, D., 2014. Automated Parameterisation for Multi- Scale Image Segmentation on Multiple Layers ISPRS Journal of Photogrammetry and Remote Sensing, 88 (100), 119–127.
5. Johnson, B., Xie, Z., 2011. Unsupervised image segmentation evaluation and refinement using a multi-scale approach. ISPRS Journal of Photogrammetry and Remote Sensing 66, 473-483.
6. Kavzoglu, T. 2017. Object-Oriented Random Forest for High Resolution Land Cover Mapping Using Quickbird-2 Imagery, Handbook of Neural Computation, pp. 607-619, pg. 658, ISBN: 9780128113196, Amsterdam: Elsevier.
7. Kavzoglu, T., Yildiz Erdemir, M., Tonbul, H., 2017. Classification of semiurban landscapes from very high resolution satellite images using a regionalized multiscale segmentation approach. Journal of Applied Remote Sensing, 11 (3), 035016.
8. Kavzoglu, T., Tonbul, H., 2018, An Experimental Comparison of Multi-Resolution Segmentation, SLIC and K- Means Clustering for Object-Based Classification of VHR Imagery. International Journal of Remote Sensing, (published online), doi.org/10.1080/01431161.2018.1506592.
9. Kim, M., Madden, M., Warner., T. A., 2009. Forest Type Mapping Using Object-specific Texture Measures from Multispectral IKONOS Imagery: Segmentation Quality and Image Classification Issues. Photogrammetric Engineering & Remote Sensing, 75 (7), 819–829.
10. Kim, M., Warner, T.A., Madden, M., Atkinson, D.S., 2011. Multi-scale GEOBIA with very high spatial resolution digital aerial imagery: scale, texture and image objects. International Journal of Remote Sensing 32 (10), 2825–2850.
11. Lenarčič, Š.A., Ritlop, K., Duric, N., Cotar, K., Oštir, K., 2015. Impact of spatial resolution on correlation between segmentation evaluation metrics and forest classification accuracy, Proceedings of SPIE – The International Society for Optical Engineering, pp. 9643,96430T.
12. Lucieer, A., Stein, A., 2002. Existential Uncertainty of Spatial Objects Segmented from Satellite Sensor Imagery. IEEE Transactions on Geoscience and Remote Sensing 40, 2518–2521.
13. Mesner, N., Oštir,K., 2014. Investigating the impact of spatial and spectral resolution of satellite images on segmentation quality. Journal of Applied Remote Sensing 8, 083696–083696.
14. Neubert, M., Herold, H., Meinel, G., 2006. Evaluation of remote sensing image segmentation quality–further results and concepts. The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, vol. XXXVI, no. 4/C42, pp. 6-11, 2006.
15. Winter, S., 2000. Location Similarity of Regions. ISPRS Journal of Photogrammetry and Remote Sensing 55 (3), 189–200.
16. Zhang, Y. J., 1996. A Survey on Evaluation Methods for Image Segmentation. Pattern Recognition 29 (8), 1335–1346.
17. Zhang, H., J., Fritts, E., Goldman, S. A., 2008. Image Segmentation Evaluation: A Survey of Unsupervised Methods. Computer Vision and Image Understanding 110, 260–280.
18. Zhang, L., Li, X., Yuan, Q., Liu, Y. 2014. Object-based approach to national land cover mapping using HJ satellite imagery. Journal of Applied Remote Sensing 8, 083686