Very high resolution (VHR) remote sensing imagery has been used for land cover classification, and it tends to a transition from land-use classification to pixel-level semantic segmentation. Inspired by the recent success of deep learning and the filter method in computer vision, this work provides a segmentation model, which designs an image segmentation neural network based on the deep residual networks and uses a guided filter to extract buildings in remote sensing imagery. Our method includes the following steps: first, the VHR remote sensing imagery is preprocessed and some hand-crafted features are calculated. Second, a designed deep network architecture is trained with the urban district remote sensing image to extract buildings at the pixel level. Third, a guided filter is employed to optimize the classification map produced by deep learning; at the same time, some salt-and-pepper noise is removed. Experimental results based on the Vaihingen and Potsdam datasets demonstrate that our method, which benefits from neural networks and guided filtering, achieves a higher overall accuracy when compared with other machine learning and deep learning methods. The method proposed shows outstanding performance in terms of the building extraction from diversified objects in the urban district.

Building Extraction in Very High Resolution Remote Sensing Imagery Using Deep Learning and Guided Filters

Joint deep learning for land cover and land use classification

Semantic segmentation of remote sensing imagery has been employed in many applications and is a key research topic for decades. With the success of deep learning methods in the field of computer vision, researchers have made a great effort to transfer their superior performance to the field of remote sensing image analysis. This paper starts with a summary of the fundamental deep neural network architectures and reviews the most recent developments of deep learning methods for semantic segmentation of remote sensing imagery including non-conventional data such as hyperspectral images and point clouds. In our review of the literature, we identified three major challenges faced by researchers and summarize the innovative development to address them. As tremendous efforts have been devoted to advancing pixel-level accuracy, the emerged deep learning methods demonstrated much-improved performance on several public data sets. As to handling the non-conventional, unstructured point cloud and rich spectral imagery, the performance of the state-of-the-art methods is, on average, inferior to that of the satellite imagery. Such a performance gap also exists in learning from small data sets. In particular, the limited non-conventional remote sensing data sets with labels is an obstacle to developing and evaluating new deep learning methods.

A review of deep learning methods for semantic segmentation of remote sensing imagery

BigData andmachine learning (ML) technologies have the potential to impactmany facets of environment andwatermanagement (EWM). BigData are information assets characterized by high volume, velocity, variety, and veracity. Fast advances in high-resolution remote sensing techniques, smart information and communication technologies, and socialmedia have contributed to the proliferation of BigData inmany EWMfields, such asweather forecasting, disastermanagement, smart water and energymanagement systems, and remote sensing. BigData brings about new opportunities for data-driven discovery in EWM, but it also requires new forms of information processing, storage, retrieval, as well as analytics.ML, a subdomain of artificial intelligence (AI), refers broadly to computer algorithms that can automatically learn fromdata.MLmay help unlock the power of BigData if properly integratedwith data analytics. Recent breakthroughs inAI and computing infrastructure have led to the fast development of powerful deep learning (DL) algorithms that can extract hierarchical features fromdata, with better predictive performance and less human intervention. Collectively BigData andML techniques have shown great potential for data-driven decisionmaking, scientific discovery, and process optimization. These technological advancesmay greatly benefit EWM, especially because (1)many EWMapplications (e.g. early floodwarning) require the capability to extract useful information from a large amount of data in autonomousmanner and in real time, (2)EWMresearches have become highlymultidisciplinary, and handling the ever increasing data volume/types using the traditional workflow is simply not an option, and last but not least, (3) the current theoretical knowledge aboutmany EWMprocesses is still incomplete, but whichmay now be complemented through data-driven discovery. A large number of applications onBigData andML have already appeared in the EWM literature in recent years. The purposes of this survey are to (1) examine the potential and benefits of data-driven research in EWM, (2) give a synopsis of key concepts and approaches in BigData andML, (3) provide a systematic review of current applications, andfinally (4) discussmajor issues and challenges, and recommend future research directions. EWM includes a broad range of research topics. Instead of attempting to survey each individual area, this review focuses on areas of nexus in EWM,with an emphasis on elucidating the potential benefits of increased data availability and predictive analytics to improving the EWMresearch.

https://iopscience.iop.org/article/10.1088/1748-9326/ab1b7d/pdf

How can Big Data and machine learning benefit environment and water management: a survey of methods, applications, and future directions

Crop type mapping without field-level labels: Random forest transfer and unsupervised clustering techniques

As a variant of Convolutional Neural Networks (CNNs) in Deep Learning, the Fully Convolutional Network (FCN) model achieved state-of-the-art performance for natural image semantic segmentation In this paper, an accurate classification approach for high resolution remote sensing imagery based on the improved FCN model is proposed Firstly, we improve the density of output class maps by introducing Atrous convolution, and secondly, we design a multi-scale network architecture by adding a skip-layer structure to make it capable for multi-resolution image classification Finally, we further refine the output class map using Conditional Random Fields (CRFs) post-processing Our classification model is trained on 70 GF-2 true color images, and tested on the other 4 GF-2 images and 3 IKONOS true color images We also employ object-oriented classification, patch-based CNN classification, and the FCN-8s approach on the same images for comparison The experiments show that compared with the existing approaches, our approach has an obvious improvement in accuracy The average precision, recall, and Kappa coefficient of our approach are 081, 078, and 083, respectively The experiments also prove that our approach has strong applicability for multi-resolution image classification

https://www.mdpi.com/2072-4292/9/5/498/pdf

Classification for High Resolution Remote Sensing Imagery Using a Fully Convolutional Network

With the acceleration of China's modernization, soil erosion caused by illegal activities in production and construction projects is becoming serious. Timely and effective monitoring and supervision is necessary. Traditional monitoring methods have many shortcomings, such as long period, limited precision, limited terrain. So it can't demonstrate the real situation of the project construction area timely and effectively. This paper attempts to apply remote sensing technique, combining the Convolution Neural Network(CNN) method, to achieve the purpose of intelligent monitoring of disturbed lands. By investigating the effect of the application of the convolutional neural network model with different training volumes, results showing that the accuracy and efficiency of the model are more and more high with the increase of training volumes. It is proved that the application of CNN in monitoring of land disturbance is effective.

Monitoring of disturbed land based on convolution neural network

The problem of color and texture distortion remains unsolved in the photorealistic style transfer task. It is mainly caused by the interference between color and texture during transferring. To address this problem, we propose a end-to-end network via adaptive filtering and channel separation. Given a pair of content image and reference image, we firstly decompose them into two structure layers through adaptive weighted least squares filter (AWLSF), which could better perceive the color structure and illumination. Then, we carry out RGB transfer in a channel separation way on the two generated structure layers. To deal with texture in a relatively independent manner, we use a module and a subtraction operation to get more complete and clear content features. Finally, we merge the color structure and texture detail into the ultimate result. We conduct solid quantitative experiments on four metrics NIQE, AG, SSIM, and PSNR, and make a user study. The experimental results demonstrate that our method is able to produce better results than previous state-of-the-art methods, and validate the effectiveness and superiority of our method.

Photorealistic Style Transfer via Adaptive Filtering and Channel Seperation

Interactive lighting editing system for single indoor low-light scene images with corresponding depth maps

In this supplementary material, we first introduce more detailed motivation clarification in Section 1, then provide the statistical analysis of distribution properties of our IH dataset in Section 2.

Gang Fu

Papers

Classification for High Resolution Remote Sensing Imagery Using a Fully Convolutional Network

Monitoring of disturbed land based on convolution neural network

Photorealistic Style Transfer via Adaptive Filtering and Channel Seperation

Interactive lighting editing system for single indoor low-light scene images with corresponding depth maps

Supplementary Materials for Deep Image-based Illumination Harmonization