Abstract:Large amounts of available training data and increasing computing power have led to the recent success of deep convolutional neural networks (CNN) on a large number of applications. In this paper, we propose an effective semantic pixel labelling using CNN features, hand-crafted features and Conditional Random Fields (CRFs). Both CNN and hand-crafted features are applied to dense image patches to produce per-pixel class probabilities. The CRF infers a labelling that smooths regions while respecting the edges pr… Show more
“…The CNN architecture used in this work is inspired on the approach presented in Paisitkriangkrai et al (2015). A scheme of this architecture is shown in Figure 1.…”
Section: Architecturementioning
confidence: 99%
“…* Corresponding author As far as our knowledge the use of CNNs for processing remotely sensed imagery is relatively recent. Particularly, CNNs have been used in remote sensing area for generating thematic maps following a pixel-based approach (Paisitkriangkrai et al, 2015;Zou et al, 2015). In a pixel-based approach, during training phase, training images are broken down into overlapping patches, where each patch is centered on a pixel which provide the class for the whole patch.…”
ABSTRACT:Recently deep learning-based methods have demonstrated excellent performance on different artificial-intelligence tasks. Even though, in the last years, several related works are found in the literature in the remote sensing field, a small percentage of them address the classification problem. These works propose schemes based on image patches to perform pixel-based image classification. Due to the typical remote sensing image size, the main drawback of these schemes is the time required by the window-sliding process implied in them. In this work, we propose a strategy to reduce the time spent on the classification of a new image through the use of superpixel segmentation. Several experiments using CNNs trained with different sizes of patches and superpixels have been performed on the ISPRS semantic labeling benchmark. Obtained results show that while the accuracy of the classification carried out by using superpixels is similar to the results generated by pixel-based approach, the expended time is dramatically decreased by means of reducing the number of elements to label.
“…The CNN architecture used in this work is inspired on the approach presented in Paisitkriangkrai et al (2015). A scheme of this architecture is shown in Figure 1.…”
Section: Architecturementioning
confidence: 99%
“…* Corresponding author As far as our knowledge the use of CNNs for processing remotely sensed imagery is relatively recent. Particularly, CNNs have been used in remote sensing area for generating thematic maps following a pixel-based approach (Paisitkriangkrai et al, 2015;Zou et al, 2015). In a pixel-based approach, during training phase, training images are broken down into overlapping patches, where each patch is centered on a pixel which provide the class for the whole patch.…”
ABSTRACT:Recently deep learning-based methods have demonstrated excellent performance on different artificial-intelligence tasks. Even though, in the last years, several related works are found in the literature in the remote sensing field, a small percentage of them address the classification problem. These works propose schemes based on image patches to perform pixel-based image classification. Due to the typical remote sensing image size, the main drawback of these schemes is the time required by the window-sliding process implied in them. In this work, we propose a strategy to reduce the time spent on the classification of a new image through the use of superpixel segmentation. Several experiments using CNNs trained with different sizes of patches and superpixels have been performed on the ISPRS semantic labeling benchmark. Obtained results show that while the accuracy of the classification carried out by using superpixels is similar to the results generated by pixel-based approach, the expended time is dramatically decreased by means of reducing the number of elements to label.
“…Therefore, using FCN for Earth Observation means we can shift from superpixel segmentation and region-based classification [4][5][6] to fully supervised semantic segmentation [7]. FCN models have been successfully applied for remote sensing data analysis, notably land cover mapping on urban areas [7,8]. For example, FCN-based models are now the state-of-the-art on the ISPRS Vaihingen Semantic Labeling dataset [9,10].…”
Like computer vision before, remote sensing has been radically changed by the introduction of deep learning and, more notably, Convolution Neural Networks. Land cover classification, object detection and scene understanding in aerial images rely more and more on deep networks to achieve new state-of-the-art results. Recent architectures such as Fully Convolutional Networks can even produce pixel level annotations for semantic mapping. In this work, we present a deep-learning based segment-before-detect method for segmentation and subsequent detection and classification of several varieties of wheeled vehicles in high resolution remote sensing images. This allows us to investigate object detection and classification on a complex dataset made up of visually similar classes, and to demonstrate the relevance of such a subclass modeling approach. Especially, we want to show that deep learning is also suitable for object-oriented analysis of Earth Observation data as effective object detection can be obtained as a byproduct of accurate semantic segmentation. First, we train a deep fully convolutional network on the ISPRS Potsdam and the NZAM/ONERA Christchurch datasets and show how the learnt semantic maps can be used to extract precise segmentation of vehicles. Then, we show that those maps are accurate enough to perform vehicle detection by simple connected component extraction. This allows us to study the repartition of vehicles in the city. Finally, we train a Convolutional Neural Network to perform vehicle classification on the VEDAI dataset, and transfer its knowledge to classify the individual vehicle instances that we detected.
“…Since the early applications to road detection back in 2010 [18], convolutional networks have been successfully used for classification and dense labeling of aerial imagery. They have defined new state-of-the-art performances and showed the re-use of cross-domain databases is possible to gain and transfer knowledge [20], [9]. New challenges will soon be addressed, such as image registration or 3D data analysis.…”
Abstract-This work shows how deep learning techniques can benefit to remote sensing. We focus on tasks which are recurrent in Earth Observation data analysis. For classification and semantic mapping of aerial images, we present various deep network architectures and show that context information and dense labeling allow to reach better performances. For estimation of normals in point clouds, combining Hough transform with convolutional networks also improves the accuracy of previous frameworks by detecting hard configurations like corners. It shows that deep learning allows to revisit remote sensing and offers promising paths for urban modeling and monitoring.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.