DEEP NEURAL NETWORKS FOR ABOVE-GROUND DETECTION IN VERY HIGH SPATIAL RESOLUTION DIGITAL ELEVATION MODELS
Deep Learning techniques have lately received increased attention for achieving state-of-the-art results in many classification problems, including various vision tasks. In this work, we implement a Deep Learning technique for classifying above-ground objects within urban environments by using a Mul...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Copernicus Publications
2015-03-01
|
| Series: | ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences |
| Online Access: | http://www.isprs-ann-photogramm-remote-sens-spatial-inf-sci.net/II-3-W4/103/2015/isprsannals-II-3-W4-103-2015.pdf |
| _version_ | 1818127854521548800 |
|---|---|
| author | D. Marmanis F. Adam M. Datcu T. Esch U. Stilla |
| author_facet | D. Marmanis F. Adam M. Datcu T. Esch U. Stilla |
| author_sort | D. Marmanis |
| collection | DOAJ |
| description | Deep Learning techniques have lately received increased attention for achieving state-of-the-art results in many classification problems, including various vision tasks. In this work, we implement a Deep Learning technique for classifying above-ground objects within urban environments by using a Multilayer Perceptron model and VHSR DEM data. In this context, we propose a novel method called M-ramp which significantly improves the classifier’s estimations by neglecting artefacts, minimizing convergence time and improving overall accuracy. We support the importance of using the M-ramp model in DEM classification by conducting a set of experiments with both quantitative and qualitative results. Precisely, we initially train our algorithm with random DEM tiles and their respective point-labels, considering less than 0.1% over the test area, depicting the city center of Munich (25 km<sup>2</sup>). Furthermore with no additional training, we classify two much larger unseen extents of the greater Munich area (424 km<sup>2</sup>) and Dongying city, China (257 km<sup>2</sup>) and evaluate their respective results for proving knowledge-transferability. Through the use of M-ramp, we were able to accelerate the convergence by a magnitude of 8 and achieve a decrease in above-ground relative error by 24.8% and 5.5% over the different datasets. |
| first_indexed | 2024-12-11T07:23:58Z |
| format | Article |
| id | doaj.art-f538232160ff40deb6d054fe2d5ca135 |
| institution | Directory Open Access Journal |
| issn | 2194-9042 2194-9050 |
| language | English |
| last_indexed | 2024-12-11T07:23:58Z |
| publishDate | 2015-03-01 |
| publisher | Copernicus Publications |
| record_format | Article |
| series | ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences |
| spelling | doaj.art-f538232160ff40deb6d054fe2d5ca1352022-12-22T01:16:00ZengCopernicus PublicationsISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences2194-90422194-90502015-03-01II-3/W410311010.5194/isprsannals-II-3-W4-103-2015DEEP NEURAL NETWORKS FOR ABOVE-GROUND DETECTION IN VERY HIGH SPATIAL RESOLUTION DIGITAL ELEVATION MODELSD. Marmanis0F. Adam1M. Datcu2T. Esch3U. Stilla4EOC, German Aerospace Center, Wessling, GermanyEOC, German Aerospace Center, Wessling, GermanyEOC, German Aerospace Center, Wessling, GermanyEOC, German Aerospace Center, Wessling, GermanyChair of Photogrammetry & Remote Sensing, Technische Universitaet Muenchen, GermanyDeep Learning techniques have lately received increased attention for achieving state-of-the-art results in many classification problems, including various vision tasks. In this work, we implement a Deep Learning technique for classifying above-ground objects within urban environments by using a Multilayer Perceptron model and VHSR DEM data. In this context, we propose a novel method called M-ramp which significantly improves the classifier’s estimations by neglecting artefacts, minimizing convergence time and improving overall accuracy. We support the importance of using the M-ramp model in DEM classification by conducting a set of experiments with both quantitative and qualitative results. Precisely, we initially train our algorithm with random DEM tiles and their respective point-labels, considering less than 0.1% over the test area, depicting the city center of Munich (25 km<sup>2</sup>). Furthermore with no additional training, we classify two much larger unseen extents of the greater Munich area (424 km<sup>2</sup>) and Dongying city, China (257 km<sup>2</sup>) and evaluate their respective results for proving knowledge-transferability. Through the use of M-ramp, we were able to accelerate the convergence by a magnitude of 8 and achieve a decrease in above-ground relative error by 24.8% and 5.5% over the different datasets.http://www.isprs-ann-photogramm-remote-sens-spatial-inf-sci.net/II-3-W4/103/2015/isprsannals-II-3-W4-103-2015.pdf |
| spellingShingle | D. Marmanis F. Adam M. Datcu T. Esch U. Stilla DEEP NEURAL NETWORKS FOR ABOVE-GROUND DETECTION IN VERY HIGH SPATIAL RESOLUTION DIGITAL ELEVATION MODELS ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences |
| title | DEEP NEURAL NETWORKS FOR ABOVE-GROUND DETECTION IN VERY HIGH SPATIAL RESOLUTION DIGITAL ELEVATION MODELS |
| title_full | DEEP NEURAL NETWORKS FOR ABOVE-GROUND DETECTION IN VERY HIGH SPATIAL RESOLUTION DIGITAL ELEVATION MODELS |
| title_fullStr | DEEP NEURAL NETWORKS FOR ABOVE-GROUND DETECTION IN VERY HIGH SPATIAL RESOLUTION DIGITAL ELEVATION MODELS |
| title_full_unstemmed | DEEP NEURAL NETWORKS FOR ABOVE-GROUND DETECTION IN VERY HIGH SPATIAL RESOLUTION DIGITAL ELEVATION MODELS |
| title_short | DEEP NEURAL NETWORKS FOR ABOVE-GROUND DETECTION IN VERY HIGH SPATIAL RESOLUTION DIGITAL ELEVATION MODELS |
| title_sort | deep neural networks for above ground detection in very high spatial resolution digital elevation models |
| url | http://www.isprs-ann-photogramm-remote-sens-spatial-inf-sci.net/II-3-W4/103/2015/isprsannals-II-3-W4-103-2015.pdf |
| work_keys_str_mv | AT dmarmanis deepneuralnetworksforabovegrounddetectioninveryhighspatialresolutiondigitalelevationmodels AT fadam deepneuralnetworksforabovegrounddetectioninveryhighspatialresolutiondigitalelevationmodels AT mdatcu deepneuralnetworksforabovegrounddetectioninveryhighspatialresolutiondigitalelevationmodels AT tesch deepneuralnetworksforabovegrounddetectioninveryhighspatialresolutiondigitalelevationmodels AT ustilla deepneuralnetworksforabovegrounddetectioninveryhighspatialresolutiondigitalelevationmodels |