Salmon, BPKleynhans, WOlivier, JCSchwegmann, CPOlding, WC2016-05-162016-05-162015-07Salmon BP, Kleynhans W, Olivier JC, Schwegmann CP, Olding WC. 2015. A multi-tier higher order Conditional Random Field for land cover classification of multi-temporal multi-spectral Landsat imagery. In: 2015 IEEE International Geoscience and Remote Sensing Symposium (IGARSS), Milan, Italy, 26-31 July 2015, pp 4372-4375.2153-6996http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=7326795http://hdl.handle.net/10204/85382015 IEEE International Geoscience and Remote Sensing Symposium (IGARSS), Milan, Italy, 26-31 July 2015. Due to copyright restrictions, the attached PDF file only contains the abstract of the full text item. For access to the full text item, please consult the publisher's websiteIn this paper the authors present a 2-tier higher order Conditional Random Field which is used for land cover classification. The Conditional Random Field is based on probabilistic messages being passed along a graph to compute efficiently the conditional probability for a land cover class. Conventionally the information is passed among direct spatial neighbors to improve classification accuracy. The inclusion of higher order descriptive structures in the graphs allow for more information to be pass along to further improve classification accuracy. Unfortunately this increases the computational cost beyond what is feasible to classify a large geographical area. In this work we investigate a spatially based cluster potential to improve classification accuracy while keeping the computational costs tractable. We also expand the typical 1-tier protograph used in conventional CRFs to a 2-tier graph to encapsulate the temporal dimension. This further improves the classification accuracy by modeling the seasonal variations experienced throughout the year. The conventional and higher order CRF are compared to a Random Forest on monthly composited Landsat images. These two CRFs are then compared to the same CRFs expanded to a 2-tier graph. An overall improvement between 2-4% is observed in our study area which is located near the city of Vryheid, South Africa.enContext awarenessGraphical ModelsImage classificationRemote SensingSatellitesStatisticsConference PresentationSalmon, B., Kleynhans, W., Olivier, J., Schwegmann, C., & Olding, W. (2015). A multi-tier higher order Conditional Random Field for land cover classification of multi-temporal multi-spectral Landsat imagery. IEEE. http://hdl.handle.net/10204/8538Salmon, BP, W Kleynhans, JC Olivier, CP Schwegmann, and WC Olding. "A multi-tier higher order Conditional Random Field for land cover classification of multi-temporal multi-spectral Landsat imagery." (2015): http://hdl.handle.net/10204/8538Salmon B, Kleynhans W, Olivier J, Schwegmann C, Olding W, A multi-tier higher order Conditional Random Field for land cover classification of multi-temporal multi-spectral Landsat imagery; IEEE; 2015. http://hdl.handle.net/10204/8538 .TY - Conference Presentation AU - Salmon, BP AU - Kleynhans, W AU - Olivier, JC AU - Schwegmann, CP AU - Olding, WC AB - In this paper the authors present a 2-tier higher order Conditional Random Field which is used for land cover classification. The Conditional Random Field is based on probabilistic messages being passed along a graph to compute efficiently the conditional probability for a land cover class. Conventionally the information is passed among direct spatial neighbors to improve classification accuracy. The inclusion of higher order descriptive structures in the graphs allow for more information to be pass along to further improve classification accuracy. Unfortunately this increases the computational cost beyond what is feasible to classify a large geographical area. In this work we investigate a spatially based cluster potential to improve classification accuracy while keeping the computational costs tractable. We also expand the typical 1-tier protograph used in conventional CRFs to a 2-tier graph to encapsulate the temporal dimension. This further improves the classification accuracy by modeling the seasonal variations experienced throughout the year. The conventional and higher order CRF are compared to a Random Forest on monthly composited Landsat images. These two CRFs are then compared to the same CRFs expanded to a 2-tier graph. An overall improvement between 2-4% is observed in our study area which is located near the city of Vryheid, South Africa. DA - 2015-07 DB - ResearchSpace DP - CSIR KW - Context awareness KW - Graphical Models KW - Image classification KW - Remote Sensing KW - Satellites KW - Statistics LK - https://researchspace.csir.co.za PY - 2015 SM - 2153-6996 T1 - A multi-tier higher order Conditional Random Field for land cover classification of multi-temporal multi-spectral Landsat imagery TI - A multi-tier higher order Conditional Random Field for land cover classification of multi-temporal multi-spectral Landsat imagery UR - http://hdl.handle.net/10204/8538 ER -