scholarly journals Classification of Very High Resolution Aerial Photos Using Spectral-Spatial Convolutional Neural Networks

2018 ◽  
Vol 2018 ◽  
pp. 1-12 ◽  
Author(s):  
Maher Ibrahim Sameen ◽  
Biswajeet Pradhan ◽  
Omar Saud Aziz
Keyword(s):  
Sensitivity Analysis ◽  
Land Cover ◽  
Ground Truth ◽  
Aerial Photographs ◽  
Dense Layer ◽  
Barren Land ◽  
Aerial Photos ◽  
Batch Normalization ◽  
Proposed Model

Classification of aerial photographs relying purely on spectral content is a challenging topic in remote sensing. A convolutional neural network (CNN) was developed to classify aerial photographs into seven land cover classes such as building, grassland, dense vegetation, waterbody, barren land, road, and shadow. The classifier utilized spectral and spatial contents of the data to maximize the accuracy of the classification process. CNN was trained from scratch with manually created ground truth samples. The architecture of the network comprised of a single convolution layer of 32 filters and a kernel size of 3 × 3, pooling size of 2 × 2, batch normalization, dropout, and a dense layer with Softmax activation. The design of the architecture and its hyperparameters were selected via sensitivity analysis and validation accuracy. The results showed that the proposed model could be effective for classifying the aerial photographs. The overall accuracy and Kappa coefficient of the best model were 0.973 and 0.967, respectively. In addition, the sensitivity analysis suggested that the use of dropout and batch normalization technique in CNN is essential to improve the generalization performance of the model. The CNN model without the techniques above achieved the worse performance, with an overall accuracy and Kappa of 0.932 and 0.922, respectively. This research shows that CNN-based models are robust for land cover classification using aerial photographs. However, the architecture and hyperparameters of these models should be carefully selected and optimized.

scholarly journals The influence of land use and land cover change on landslide susceptibility: a case study in Zhushan Town, Xuan'en County (Hubei, China)

2019 ◽  
Vol 19 (10) ◽  
pp. 2207-2228 ◽  
Author(s):  
Lixia Chen ◽  
Zizheng Guo ◽  
Kunlong Yin ◽  
Dhruba Pikha Shrestha ◽  
Shikuan Jin
Keyword(s):  
Land Use ◽  
Land Cover ◽  
Land Cover Change ◽  
Landslide Susceptibility ◽  
Regional Scale ◽  
Aerial Photographs ◽  
Time Interval ◽  
Human Engineering ◽  
Classification Of Images

Abstract. Land use and land cover change can increase or decrease landslide susceptibility (LS) in the mountainous areas. In the hilly and mountainous part of southwestern China, land use and land cover change (LUCC) has taken place in the last decades due to infrastructure development and rapid economic activities. This development and activities can worsen the slope susceptible to sliding due to mostly the cutting of slopes. This study, taking Zhushan Town, Xuan'en County, as the study area, aims to evaluate the influence of land use and land cover change on landslide susceptibility at a regional scale. Spatial distribution of landslides was determined in terms of visual interpretation of aerial photographs and remote sensing images, supported by field surveys. Two types of land use and land cover (LUC) maps, with a time interval covering 21 years (1992–2013), were prepared: the first was obtained by the neural net classification of images acquired in 1992 and the second by the object-oriented classification of images in 2002 and 2013. Landslide-susceptible areas were analyzed using the logistic regression model (LRM) in which six influencing factors were chosen as the landslide susceptibility indices. In addition, the hydrologic analysis method was applied to optimize the partitioning of the terrain. The results indicated that the LUCC in the region was mainly the transformation from the grassland and arable land to the forest land, which is increased by 34.3 %. An increase of 1.9 % is shown in the area where human engineering activities concentrate. The comparison of landslide susceptibility maps among different periods revealed that human engineering activities were the most important factor in increasing LS in this region. Such results emphasize the requirement of a reasonable land use planning activity process.

scholarly journals LAND USE AND LAND COVER CLASSIFICATION OF SENTINEL 2-A: ST PETERSBURG CASE STUDY

2019 ◽  
Vol XLII-1/W2 ◽  
pp. 13-16 ◽  
Author(s):  
M. Cavur ◽  
H. S. Duzgun ◽  
S. Kemec ◽  
D. C. Demirkan
Keyword(s):  
Land Use ◽  
Land Cover ◽  
Accuracy Assessment ◽  
Ground Truth ◽  
Remotely Sensed Data ◽  
Vector Machines ◽  
Kappa Value ◽  
Sentinel 2

<p><strong>Abstract.</strong> Land use and land cover (LULC) maps in many areas have been used by companies, government offices, municipalities, and ministries. Accurate classification for LULC using remotely sensed data requires State of Art classification methods. The SNAP free software and ArcGIS Desktop were used for analysis and report. In this study, the optical Sentinel-2 images were used. In order to analyze the data, an object-oriented method was applied: Supported Vector Machines (SVM). An accuracy assessment is also applied to the classified results based on the ground truth points or known reference pixels. The overall classification accuracy of 83,64% with the kappa value of 0.802 was achieved using SVM. The study indicated that of SVM algorithms, the proposed framework on Sentinel-2 imagery results is satisfactory for LULC maps.</p>

scholarly journals The influence of land use and land cover change on landslide susceptibility: A case study in Zhushan Town, Xuanen County (Hubei, China)

2019 ◽  
Author(s):  
Lixia Chen ◽  
Zizheng Guo ◽  
Kunlong Yin ◽  
Dhruba Pikha Shrestha ◽  
Shikuan Jin
Keyword(s):  
Remote Sensing ◽  
Land Use ◽  
Land Cover ◽  
Land Cover Change ◽  
Landslide Susceptibility ◽  
Aerial Photographs ◽  
Time Interval ◽  
Human Engineering ◽  
Classification Of Images

Abstract. Land use and land cover change can have effect on the land by increasing/decreasing landslide susceptibility (LS) in the mountainous areas. In the southwestern hilly and mountainous part of China, land use and land cover change (LUCC) has been taking place in the recent past due to infrastructure development and increase in economic activities. These development activities can also bring negative effects: the sloping area may become susceptible to landsliding due to undercutting of slopes. The study aims at evaluating the influence of land use and land cover change on landslide susceptibility at regional scale, based on the application of Geographic Information System (GIS) and Remote Sensing (RS) technologies. Specific objective is to answer the question: which land cover/land use change poses the highest risk so that mitigation measures can be implemented in time? The Zhushan Town, Xuanen County in the southwest of China was taken as the study area and the spatial distribution of landslides was determined from visual interpretation of aerial photographs and remote sensing images, as well as field survey. Two types of land use/land cover (LUC) maps, with a time interval covering 21 years (1992–2013), were prepared: the first was obtained through the neural net classification of images acquired in 1992, the second through the object-oriented classification of images in 2002 and 2013. Landslide susceptible areas were analyzed using logistic regression models. In this process, six landslide influencing factors were chosen as the landslide susceptibility indices. Moreover, we applied a hydrologic analysis method achieving slope unit (SU) delineation to optimize the partitioning of the terrain. The results indicate that the LUCC in the region was mainly the transformation from the grassland and arable land to the forest land and the human engineering activities land (HEAL). The areas of these two kind of LUC increased by 34.3 % and 1.9 %, respectively. The comparison of landslide susceptibility maps in various periods revealed that human engineering activities was the most important factor to increase LS in this region. Such results underline that a more reasonable land use planning in the urbanization process is necessary.

scholarly journals Land Use and Land Cover Classification from ETM Sensor Data : A Case Study from Tamakoshi River Basin of Nepal

2014 ◽  
Vol XL-8 ◽  
pp. 943-948
Author(s):  
U. S. Shrestha
Keyword(s):  
Land Use ◽  
Land Cover ◽  
Image Classification ◽  
Land Cover Classification ◽  
Sensor Data ◽  
Classification Methods ◽  
Barren Land ◽  
Grass Land ◽  
Mountain Watershed

The mountain watershed of Nepal is highly rugged, inaccessible and difficult for acquiring field data. The application of ETM sensor Data Sat satellite image of 30 meter pixel resolutions has been used for land use and land cover classification of Tamakoshi River Basin (TRB) of Nepal. The paper tries to examine the strength of image classification methods in derivation of land use and land classification. Supervised digital image classification techniques was used for examination the thematic classification. Field verification, Google earth image, aerial photographs, topographical sheet and GPS locations were used for land use and land cover type classification, selecting training samples and assessing accuracy of classification results. Six major land use and land cover types: forest land, water bodies, bush/grass land, barren land, snow land and agricultural land was extracted using the method. Moreover, there is spatial variation of statistics of classified land uses and land cover types depending upon the classification methods. <br><br> The image data revealed that the major portion of the surface area is covered by unclassified bush and grass land covering 34.62 per cent followed by barren land (28 per cent). The knowledge derived from supervised classification was applied for the study. The result based on the field survey of the area during July 2014 also verifies the same result. So image classification is found more reliable in land use and land cover classification of mountain watershed of Nepal.

scholarly journals Supervised and Semi-Supervised Self-Organizing Maps for Regression and Classification Focusing on Hyperspectral Data

2019 ◽  
Vol 12 (1) ◽  
pp. 7 ◽  
Author(s):  
Felix M. Riese ◽  
Sina Keller ◽  
Stefan Hinz
Keyword(s):  
Land Cover ◽  
Reference Data ◽  
Ground Truth ◽  
Hyperspectral Data ◽  
Physical Parameters ◽  
Learning Approaches ◽  
Self Organizing Maps ◽  
Python Package ◽  
Self Organizing

Machine learning approaches are valuable methods in hyperspectral remote sensing, especially for the classification of land cover or for the regression of physical parameters. While the recording of hyperspectral data has become affordable with innovative technologies, the acquisition of reference data (ground truth) has remained expensive and time-consuming. There is a need for methodological approaches that can handle datasets with significantly more hyperspectral input data than reference data. We introduce the Supervised Self-organizing Maps (SuSi) framework, which can perform unsupervised, supervised and semi-supervised classification as well as regression on high-dimensional data. The methodology of the SuSi framework is presented and compared to other frameworks. Its different parts are evaluated on two hyperspectral datasets. The results of the evaluations can be summarized in four major findings: (1) The supervised and semi-Supervised Self-organizing Maps (SOM) outperform random forest in the regression of soil moisture. (2) In the classification of land cover, the supervised and semi-supervised SOM reveal great potential. (3) The unsupervised SOM is a valuable tool to understand the data. (4) The SuSi framework is versatile, flexible, and easy to use. The SuSi framework is provided as an open-source Python package on GitHub.

scholarly journals Assessing Land Use and Land Cover Change in River Beas Floodplain, Punjab

2020 ◽  
Vol 15 (1) ◽  
pp. 52-58
Author(s):  
Gursewak Singh Brar ◽  
Vishwa B.S. Chandel ◽  
Karanjot Kaur Brar ◽  
◽  
◽  
...  
Keyword(s):  
Land Use ◽  
Land Cover ◽  
Land Use Changes ◽  
River Channels ◽  
Human Environment ◽  
Barren Land ◽  
Natural Land ◽  
Multi Temporal ◽  
River Beas

Floodplains are the most fragile ecosystems of the world which attracted the humans since the dawn of civilizations. Due to their resource enrichment, these remained center of attraction to fulfill the socio-economic needs of people. As a result, the natural land cover of these floodplains are under the influence of human induced activities. River Beas Floodplain of Punjab has also witnessed such changes. Human intervention in these landscapes has depleted natural wealth and has altered its land use. Construction of upstream dam and artificial embankments and diversion of water through canals further paved the ways for intensification of land use changes. The outcome of these human actions is that wetlands, barren land, and river channels has reduced. On the other hand, agriculture and settlements recorded a sharp increase in recent decades. The growth of agricultural area and human settlements are putting pressure on the natural resources and depleting the human environment relationship in the floodplain. This study utilized multi-temporal satellite data from Landsat for the classification of land use and land cover.

scholarly journals Use of Remote Sensing and Geographic Information System for the Classification of Agricultural Land Uses and Land Cover in the Al-Sad Al-Adhim sub District – Iraq

2018 ◽  
Vol 225 (2) ◽  
pp. 245-273
Author(s):  
Assist. Prof. Dr. Saleem Y. Jamal
Keyword(s):  
Remote Sensing ◽  
Land Use ◽  
Information System ◽  
Geographic Information System ◽  
Land Cover ◽  
Agricultural Land ◽  
Geographic Information ◽  
Land Uses ◽  
Barren Land

     Land use refers to the human activity associated with a particular area of land. The land cover refers to the pattern of appearances located on the surface of the earth. Survey, inventory, monitoring and classification of land use and land cover are a fundamental step in the land use planning process, in evaluating and comparing alternatives and in choosing the best and sustainable use of land for development, accomplishment economic and social well-being. Remote sensing and Geographic Information System provided advantages that conventional methods could not provide for surveys and monitoring of natural and human resources, and classification of agricultural land uses and land cover in the area of the Al-Sad Al-Adhim sub District – Iraq. Depending on the Anderson system and others to classify land uses and land cover, through the integration of digital interpretation with the use of Digital Image Processing (ERDAS IMAGINE) software, and visual interpretation using ArcGIS software. Classification of agricultural land use and land cover up to the third level, with over all accuracy of the map 90%. the percentage distribution of the areas shows that the agricultural lands ranked first and occupy 52%, then grassland occupies 19%, barren land is occupied 17%, urban areas and built up occupy 9% water is ranked last occupy 3% of the total area of the study area.

scholarly journals Semantic Boosting: Enhancing Deep Learning Based LULC Classification

2021 ◽  
Vol 13 (16) ◽  
pp. 3197
Author(s):  
Marvin Mc Mc Cutchan ◽  
Alexis J. Comber ◽  
Ioannis Giannopoulos ◽  
Manuela Canestrini
Keyword(s):  
Land Use ◽  
Deep Learning ◽  
Land Cover ◽  
Information Science ◽  
Ground Truth ◽  
Vector Data ◽  
Remotely Sensed Imagery ◽  
The Impact ◽  
Sentinel 2

The classification of land use and land cover (LULC) is a well-studied task within the domain of remote sensing and geographic information science. It traditionally relies on remotely sensed imagery and therefore models land cover classes with respect to their electromagnetic reflectances, aggregated in pixels. This paper introduces a methodology which enables the inclusion of geographical object semantics (from vector data) into the LULC classification procedure. As such, information on the types of geographic objects (e.g., Shop, Church, Peak, etc.) can improve LULC classification accuracy. In this paper, we demonstrate how semantics can be fused with imagery to classify LULC. Three experiments were performed to explore and highlight the impact and potential of semantics for this task. In each experiment CORINE LULC data was used as a ground truth and predicted using imagery from Sentinel-2 and semantics from LinkedGeoData using deep learning. Our results reveal that LULC can be classified from semantics only and that fusing semantics with imagery—Semantic Boosting—improved the classification with significantly higher LULC accuracies. The results show that some LULC classes are better predicted using only semantics, others with just imagery, and importantly much of the improvement was due to the ability to separate similar land use classes. A number of key considerations are discussed.

scholarly journals Implementation of Transfer Learning Using VGG16 on Fruit Ripeness Detection

2021 ◽  
Vol 13 (2) ◽  
pp. 52-61
Author(s):  
Jasman Pardede ◽  
◽  
Benhard Sitohang ◽  
Saiful Akbar ◽  
Masayu Leylia Khodra
Keyword(s):  
Transfer Learning ◽  
Activation Function ◽  
Experimental Results ◽  
Feature Descriptor ◽  
Dense Layer ◽  
Batch Normalization ◽  
Learning Techniques ◽  
Real World Applications

In previous studies, researchers have determined the classification of fruit ripeness using the feature descriptor using color features (RGB, GSL, HSV, and L * a * b *). However, the performance from the experimental results obtained still yields results that are less than the maximum, viz the maximal accuracy is only 76%. Today, transfer learning techniques have been applied successfully in many real-world applications. For this reason, researchers propose transfer learning techniques using the VGG16 model. The proposed architecture uses VGG16 without the top layer. The top layer of the VGG16 replaced by adding a Multilayer Perceptron (MLP) block. The MLP block contains Flatten layer, a Dense layer, and Regularizes. The output of the MLP block uses the softmax activation function. There are three Regularizes that considered in the MLP block namely Dropout, Batch Normalization, and Regularizes kernels. The Regularizes selected are intended to reduce overfitting. The proposed architecture conducted on a fruit ripeness dataset that was created by researchers. Based on the experimental results found that the performance of the proposed architecture has better performance. Determination of the type of Regularizes is very influential on system performance. The best performance obtained on the MLP block that has Dropout 0.5 with increased accuracy reaching 18.42%. The Batch Normalization and the Regularizes kernels performance increased the accuracy amount of 10.52% and 2.63%, respectively. This study shows that the performance of deep learning using transfer learning always gets better performance than using machine learning with traditional feature extraction to determines fruit ripeness detection. This study gives also declaring that Dropout is the best technique to reduce overfitting in transfer learning.

Classification of vegetation in Cucphuong National Park with the aim of large-scale mapping, Vietnam

1996 ◽  
pp. 64-67 ◽  
Author(s):  
Nguen Nghia Thin ◽  
Nguen Ba Thu ◽  
Tran Van Thuy
Keyword(s):  
National Park ◽  
Large Scale ◽  
Satellite Images ◽  
Field Work ◽  
Aerial Photographs ◽  
Vegetation Map ◽  
Dense Forest ◽  
Broad Leaved Forest ◽  
Leaved Forest

The tropical seasonal rainy evergreen broad-leaved forest vegetation of the Cucphoung National Park has been classified and the distribution of plant communities has been shown on the map using the relations of vegetation to geology, geomorphology and pedology. The method of vegetation mapping includes: 1) the identifying of vegetation types in the remote-sensed materials (aerial photographs and satellite images); 2) field work to compile the interpretation keys and to characterize all the communities of a study area; 3) compilation of the final vegetation map using the combined information. In the classification presented a number of different level vegetation units have been identified: formation classes (3), formation sub-classes (3), formation groups (3), formations (4), subformations (10) and communities (19). Communities have been taken as mapping units. So in the vegetation map of the National Park 19 vegetation categories has been shown altogether, among them 13 are natural primary communities, and 6 are the secondary, anthropogenic ones. The secondary succession goes through 3 main stages: grassland herbaceous xerophytic vegetation, xerophytic scrub, dense forest.

Sign in / Sign up

Export Citation Format

Share Document