Deep Learning Based Domain Adaptation with Data Fusion for Aerial Image Data Analysis

2021 ◽  
pp. 118-133
Author(s):  
Jingyang Lu ◽  
Chenggang Yu ◽  
Erik Blasch ◽  
Roman Ilin ◽  
Hua-mei Chen ◽  
...  
Keyword(s):  
Deep Learning ◽  
Data Analysis ◽  
Data Fusion ◽  
Domain Adaptation ◽  
Image Data ◽  
Aerial Image

scholarly journals A Hierarchical Deep-Learning Approach for Rapid Windthrow Detection on PlanetScope and High-Resolution Aerial Image Data

2020 ◽  
Vol 12 (13) ◽  
pp. 2121 ◽  
Author(s):  
Wolfgang Deigele ◽  
Melanie Brandmeier ◽  
Christoph Straub
Keyword(s):  
Deep Learning ◽  
High Resolution ◽  
Transfer Learning ◽  
Image Data ◽  
Aerial Image ◽  
Forest Damage ◽  
Computational Time ◽  
Storm Event ◽  
High Temporal Resolution ◽  
State Institute

Forest damage due to storms causes economic loss and requires a fast response to prevent further damage such as bark beetle infestations. By using Convolutional Neural Networks (CNNs) in conjunction with a GIS, we aim at completely streamlining the detection and mapping process for forest agencies. We developed and tested different CNNs for rapid windthrow detection based on PlanetScope satellite data and high-resolution aerial image data. Depending on the meteorological situation after the storm, PlanetScope data might be rapidly available due to its high temporal resolution, while the acquisition of high-resolution airborne data often takes weeks to a month and is, therefore, used in a second step for more detailed mapping. The study area is located in Bavaria, Germany (ca. 165 km2), and labels for damaged areas were provided by the Bavarian State Institute of Forestry (LWF). Modifications of a U-Net architecture were compared to other approaches using transfer learning (e.g., VGG19) to find the most efficient architecture for the task on both datasets while keeping the computational time low. A custom implementation of U-Net proved to be more accurate than transfer learning, especially on medium (3 m) resolution PlanetScope imagery (intersection over union score (IoU) 0.55) where transfer learning completely failed. Results for transfer learning based on VGG19 on high-resolution aerial image data are comparable to results from the custom U-Net architecture (IoU 0.76 vs. 0.73). When using both architectures on a dataset from a different area (located in Hesse, Germany), however, we find that the custom implementations have problems generalizing on aerial image data while VGG19 still detects most damage in these images. For PlanetScope data, VGG19 again fails while U-Net achieves reasonable mappings. Results highlight the potential of Deep Learning algorithms to detect damaged areas with an IoU of 0.73 on airborne data and 0.55 on Planet Dove data. The proposed workflow with complete integration into ArcGIS is well-suited for rapid first assessments after a storm event that allows for better planning of the flight campaign followed by detailed mapping in a second stage.

scholarly journals Batik Classification using Deep Learning

2019 ◽  
Vol 8 (4) ◽  
pp. 11416-11421
Keyword(s):  
Neural Network ◽  
Deep Learning ◽  
Data Analysis ◽  
Convolutional Neural Network ◽  
Image Data ◽  
High Accuracy ◽  
Learning Method ◽  
Philosophy Of Life ◽  
Proposed Model

Batik is one of the Indonesian cultural heritages that has been recognized by the global community. Indonesian batik has a vast diversity in motifs that illustrate the philosophy of life, the ancestral heritage and also reflects the origin of batik itself. Because of the manybatik motifs, problems arise in determining the type of batik itself. Therefore, we need a classification method that can classify various batik motifs automatically based on the batik images. The technique of image classification that is used widely now is deep learning method. This technique has been proven of its capacity in identifying images in high accuracy. Architecture that is widely used for the image data analysis is Convolutional Neural Network (CNN) because this architecture is able to detect and recognize objects in an image. This workproposes to use the method of CNN and VGG architecture that have been modified to overcome the problems of classification of the batik motifs. Experiments of using 2.448 batik images from 5 classes of batik motifs showed that the proposed model has successfully achieved an accuracy of 96.30%.

Estimating damage size and remaining useful life in degraded structures using deep learning-based multi-source data fusion

2019 ◽  
Vol 19 (5) ◽  
pp. 1542-1559 ◽  
Author(s):  
Amin Aria ◽  
Enrique Lopez Droguett ◽  
Shapour Azarm ◽  
Mohammad Modarres
Keyword(s):  
Deep Learning ◽  
Data Fusion ◽  
Short Term Memory ◽  
Image Data ◽  
Crack Size ◽  
Remaining Useful Life ◽  
Life Estimation ◽  
Temporal Correlations ◽  
Segmentation Approach ◽  
Useful Life

In this article, a new deep learning-based approach for online estimation of damage size and remaining useful life of structures is presented. The proposed approach consists of three modules. In the first module, a long short-term memory regression model is used to construct a sensor-based estimation of the damage size where different ranges of temporal correlations are considered for their effects on the accuracy of the damage size estimations. In the second module, a convolutional neural network semantic image segmentation approach is used to construct automated damage size estimations in which a pixel-wise classification is carried out on images of the damaged areas. Using physics-of-failure relations, frequency mismatches associated with sensor- and image-based size estimations are resolved. Finally, in the third module, damage size estimations obtained by the first two modules are fused together for an online remaining useful life estimation of the structure. Performance of the proposed approach is evaluated using sensor and image data obtained from a set of fatigue crack experiments performed on aluminum alloy 7075-T6 specimens. It is shown that using acoustic emission signals obtained from sensors and microscopic images in these experiments, the damage size estimations obtained from the proposed data fusion approach have higher accuracy than the sensor-based and higher frequency than the image-based estimations. Moreover, the accuracy of the data fusion estimations is found to be more than that of image-based estimations for the experiment with the largest sensor dataset. Based on the results obtained, it is concluded that the consideration of longer temporal correlations can lead to improvements in the accuracy of crack size estimations and, thus, a better remaining useful life estimation for structures.

scholarly journals Hybrid Technique for Medical Data Classification using Multi-Layer Perceptron with NB Classifier

2019 ◽  
Vol 8 (12) ◽  
pp. 2627-2632
Keyword(s):  
Deep Learning ◽  
Data Analysis ◽  
Image Data ◽  
Numerical Data ◽  
Data Classification ◽  
Heterogeneous Data ◽  
Medical Data ◽  
Hybrid Technique ◽  
Text Data ◽  
Medical Data Classification

Medical data analysis gains more interest from the last decade due to its significance advantages. Medical data is a heterogeneous data, which is the combination of text data, numeric data and image data. For to analyze such heterogeneous data traditional data analysis mechanisms are inefficient. To handle this heterogeneous data deep learning is obvious choice. Deep learning is able to handle text, numeric and image data more efficiently than traditional data mining techniques. In this paper we proposed a deep learning based multilayer perceptron to analysis medical data. This method independently address the text data, image data and numerical data and combinable made medical data classification

Deep Learning Approaches for Whiteboard Image Quality Enhancement

2019 ◽  
Vol 2019 (1) ◽  
pp. 360-368
Author(s):  
Mekides Assefa Abebe ◽  
Jon Yngve Hardeberg
Keyword(s):  
Deep Learning ◽  
Image Quality ◽  
Image Data ◽  
Quality Enhancement ◽  
Network Architectures ◽  
Learning Approaches ◽  
Data Set ◽  
Image Quality Enhancement ◽  
Processing Techniques ◽  
White Balancing

Different whiteboard image degradations highly reduce the legibility of pen-stroke content as well as the overall quality of the images. Consequently, different researchers addressed the problem through different image enhancement techniques. Most of the state-of-the-art approaches applied common image processing techniques such as background foreground segmentation, text extraction, contrast and color enhancements and white balancing. However, such types of conventional enhancement methods are incapable of recovering severely degraded pen-stroke contents and produce artifacts in the presence of complex pen-stroke illustrations. In order to surmount such problems, the authors have proposed a deep learning based solution. They have contributed a new whiteboard image data set and adopted two deep convolutional neural network architectures for whiteboard image quality enhancement applications. Their different evaluations of the trained models demonstrated their superior performances over the conventional methods.

scholarly journals Fig Plant Segmentation from Aerial Images Using a Deep Convolutional Encoder-Decoder Network

2019 ◽  
Vol 11 (10) ◽  
pp. 1157 ◽  
Author(s):  
Jorge Fuentes-Pacheco ◽  
Juan Torres-Olivares ◽  
Edgar Roman-Rangel ◽  
Salvador Cervantes ◽  
Porfirio Juarez-Lopez ◽  
...  
Keyword(s):  
Precision Agriculture ◽  
Image Data ◽  
Ground Truth ◽  
Aerial Images ◽  
Aerial Image ◽  
Data Set ◽  
Visual Appearance ◽  
Aerial Robots ◽  
Lighting Conditions ◽  
Convolutional Encoder

Crop segmentation is an important task in Precision Agriculture, where the use of aerial robots with an on-board camera has contributed to the development of new solution alternatives. We address the problem of fig plant segmentation in top-view RGB (Red-Green-Blue) images of a crop grown under open-field difficult circumstances of complex lighting conditions and non-ideal crop maintenance practices defined by local farmers. We present a Convolutional Neural Network (CNN) with an encoder-decoder architecture that classifies each pixel as crop or non-crop using only raw colour images as input. Our approach achieves a mean accuracy of 93.85% despite the complexity of the background and a highly variable visual appearance of the leaves. We make available our CNN code to the research community, as well as the aerial image data set and a hand-made ground truth segmentation with pixel precision to facilitate the comparison among different algorithms.

Image data analysis and classification in marketing

2012 ◽  
Vol 6 (4) ◽  
pp. 253-276 ◽  
Author(s):  
Daniel Baier ◽  
Ines Daniel ◽  
Sarah Frost ◽  
Robert Naundorf
Keyword(s):  
Data Analysis ◽  
Image Data
Sign in / Sign up

Export Citation Format

Share Document