Generating Elevation Surface from a Single RGB Remotely Sensed Image Using Deep Learning

Panagiotou, Emmanouil; Chochlakis, Georgios; Γραμματικόπουλος, Λάζαρος; Charou, Eleni

doi:10.3390/rs12122002

Cited by 16 publications

(10 citation statements)

References 79 publications

(105 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Assessment methods are also necessary for quantifying model sensitivity to hyperparameter settings (e.g., Li and Hsu [73]), such as architecture augmentations; loss metric, optimization algorithm, or learning rate used, and learning rate scheduling applied. Accuracy assessment is also important when studying the impact of training data quantity (i.e., ablation studies), quality, and/or augmentation (e.g., [38,68,74,75]).…”

Section: Deep Learning Accuracy Assessment Example Use Casesmentioning

confidence: 99%

Accuracy Assessment in Convolutional Neural Network-Based Deep Learning Remote Sensing Studies—Part 1: Literature Review

2021

View full text Add to dashboard Cite

Convolutional neural network (CNN)-based deep learning (DL) is a powerful, recently developed image classification approach. With origins in the computer vision and image processing communities, the accuracy assessment methods developed for CNN-based DL use a wide range of metrics that may be unfamiliar to the remote sensing (RS) community. To explore the differences between traditional RS and DL RS methods, we surveyed a random selection of 100 papers from the RS DL literature. The results show that RS DL studies have largely abandoned traditional RS accuracy assessment terminology, though some of the accuracy measures typically used in DL papers, most notably precision and recall, have direct equivalents in traditional RS terminology. Some of the DL accuracy terms have multiple names, or are equivalent to another measure. In our sample, DL studies only rarely reported a complete confusion matrix, and when they did so, it was even more rare that the confusion matrix estimated population properties. On the other hand, some DL studies are increasingly paying attention to the role of class prevalence in designing accuracy assessment approaches. DL studies that evaluate the decision boundary threshold over a range of values tend to use the precision-recall (P-R) curve, the associated area under the curve (AUC) measures of average precision (AP) and mean average precision (mAP), rather than the traditional receiver operating characteristic (ROC) curve and its AUC. DL studies are also notable for testing the generalization of their models on entirely new datasets, including data from new areas, new acquisition times, or even new sensors.

show abstract

Section: Deep Learning Accuracy Assessment Example Use Casesmentioning

confidence: 99%

Accuracy Assessment in Convolutional Neural Network-Based Deep Learning Remote Sensing Studies—Part 1: Literature Review

2021

View full text Add to dashboard Cite

show abstract

“…however, the recorded data were obtained from a short highway segment and the use of velocity direction with respect to the reference frame needs further investigation. Moreover, in the work done by [144], a comprehensive deep learning methodology was proposed to generate an absolute or relative point cloud estimation of a digital elevation model (DEM) given a single satellite or drone image for a wide range of applications and disciplines such as 3D flight planning, autonomous driving, and satellite navigation. Shan et al [145] also formed a new method based on a machine learning approach to collect and share data among drones and other aircraft, analyze data and establish models, and capture more detailed characteristics about drone communications, which is useful for avoiding hazardous conditions.…”

Section: Quad-rotor Systemsmentioning

confidence: 99%

Control Strategies and Novel Techniques for Autonomous Rotorcraft Unmanned Aerial Vehicles: A Review

2020

View full text Add to dashboard Cite

This paper presents a review of the various control strategies that have been conducted to address and resolve several challenges for a particular category of unmanned aerial vehicles (UAVs), the emphasis of which is on the rotorcraft or rotary-wing systems. Initially, a brief overview of the important relevant definitions, configurations, components, advantages/disadvantages, and applications of the UAVs is first introduced in general, encompassing a wide spectrum of the flying machines. Subsequently, the focus is more on the two most common and versatile rotorcraft UAVs, namely, the twin-rotor and quadrotor systems. Starting with a brief background on the dual-rotor helicopter and a quadcopter, the full detailed mathematical dynamic model of each system is derived based on the Euler-Lagrange and Newton-Euler methods, considering a number of assumptions and considerations. Then, a state-of-the-art review of the diverse control strategies for controlling the rotorcraft systems with conceivable solutions when the systems are subjected to the different impediments is demonstrated. To counter some of these limitations and adverse operating/loading conditions in the UAVs, several innovative control techniques are particularly highlighted, and their performance are duly analyzed, discussed, and compared. The applied control techniques are deemed to produce a useful contribution to their successful implementation in the wake of varied constraints and demanding environments that result in a degree of robustness and efficacy. Some of the off-the-shelf developments in the rotorcraft systems for research and commercial applications are also presented.

show abstract

“…The PatchGAN architecture can be seen in Figure 6. The task of predicting plausible DEMs for input remotely sensed imagery , as well as, model evaluation and accuracy have been addressed thoroughly in our previous work [11].…”

Section: Conditional Generative Adversarial Network For Elevationmentioning

confidence: 99%

Procedural 3D Terrain Generation using Generative Adversarial Networks

Panagiotou¹,

Charou²

2020

Preprint

Self Cite

View full text Add to dashboard Cite

Generating Elevation Surface from a Single RGB Remotely Sensed Image Using Deep Learning

Cited by 16 publications

References 79 publications

Accuracy Assessment in Convolutional Neural Network-Based Deep Learning Remote Sensing Studies—Part 1: Literature Review

Accuracy Assessment in Convolutional Neural Network-Based Deep Learning Remote Sensing Studies—Part 1: Literature Review

Control Strategies and Novel Techniques for Autonomous Rotorcraft Unmanned Aerial Vehicles: A Review

Procedural 3D Terrain Generation using Generative Adversarial Networks

Contact Info

Product

Resources

About