Improved VHR Urban Area Mapping Exploiting Object Boundaries

Gamba, Paolo; Acqua, F. Dell'; Lisini, Gianni; Trianni, G.

doi:10.1109/tgrs.2007.899811

Cited by 67 publications

(27 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…In Sirmacek (2011) various local features are detected and used in detection of urban area using variable kernel based density estimation method. Gamba (2007) utilize the boundary information for urban area mapping. The boundary and non-boundary pixels are classified using neural network and Markov Random Field (MRF) classifiers respectively and the results are combined using decision fusion.…”

Section: Introductionmentioning

confidence: 99%

Extraction of Built-Up Areas Using Convolutional Neural Networks and Transfer Learning From Sentinel-2 Satellite Images

Bramhe¹,

Ghosh²,

Garg³

2018

Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci.

View full text Add to dashboard Cite

ABSTRACT:With rapid globalization, the extent of built-up areas is continuously increasing. Extraction of features for classifying built-up areas that are more robust and abstract is a leading research topic from past many years. Although, various studies have been carried out where spatial information along with spectral features has been utilized to enhance the accuracy of classification. Still, these feature extraction techniques require a large number of user-specific parameters and generally application specific. On the other hand, recently introduced Deep Learning (DL) techniques requires less number of parameters to represent more abstract aspects of the data without any manual effort. Since, it is difficult to acquire high-resolution datasets for applications that require large scale monitoring of areas. Therefore, in this study Sentinel-2 image has been used for built-up areas extraction. In this work, pre-trained Convolutional Neural Networks (ConvNets) i.e. Inception v3 and VGGNet are employed for transfer learning. Since these networks are trained on generic images of ImageNet dataset which are having very different characteristics from satellite images. Therefore, weights of networks are fine-tuned using data derived from Sentinel-2 images. To compare the accuracies with existing shallow networks, two state of art classifiers i.e. Gaussian Support Vector Machine (SVM) and Back-Propagation Neural Network (BP-NN) are also implemented. Both SVM and BP-NN gives 84.31% and 82.86% overall accuracies respectively. Inception-v3 and VGGNet gives 89.43% of overall accuracy using fine-tuned VGGNet and 92.10% when using Inception-v3. The results indicate high accuracy of proposed fine-tuned ConvNets on a 4-channel Sentinel-2 dataset for built-up area extraction.

show abstract

Section: Introductionmentioning

confidence: 99%

Extraction of Built-Up Areas Using Convolutional Neural Networks and Transfer Learning From Sentinel-2 Satellite Images

Bramhe¹,

Ghosh²,

Garg³

2018

Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci.

View full text Add to dashboard Cite

show abstract

“…A lot of previous research used analytical approach, which focused on pixel-or object based classification. It extracted spectral, texture, and geometrical attributes [8][9][10][11][12]. Nevertheless, the attribute is only used in a certain environment so it just produce less data representation.…”

Section: Introductionmentioning

confidence: 99%

Sparse Coding-Based Method Comparison For Land-Use Classification

Arsa

Jati

Hilman

2017

Jurnal Ilmu Komputer dan Informasi

View full text Add to dashboard Cite

Land-use classification utilize high-resolution remote sensing image. The image is utilized for improving the classification problem. Nonetheless, in other side, the problem becomes more challenging cause the image is too complex. We have to represent the image appropriately. On of the common method to deal with it is Bag of Visual Word (BOVW). The method needs a coding process to get the final data interpretation. There are many methods to do coding such as Hard Quantization Coding (HQ), Sparse Coding (SC), and Locality-constrained Linear Coding (LCC). However, that coding methods use a different assumption. Therefore, we have to compare the result of each coding method. The coding method affects classification accuracy. The best coding method will produce the better classification result. Dataset UC Merced consisted 21 classes is used in this research. The experiment result shows that LCC got better performance / accuracy than SC and HQ. LCC method got 86.48 % accuracy. Furthermore, LCC also got the best performance on various number of training data for each class. Keywords: Land-use classification, high-resolution remote sensing image, Bag of Visual Word (BOVW), Sparse Coding (SC), Hard Quantization Coding (HQ) AbstrakKlasifikasi penggunaan lahan memanfaatkan gambar penginderaan jauh beresolusi tinggi. Citra digunakan untuk memperbaiki masalah klasifikasi. Meski begitu, di sisi lain, masalahnya menjadi lebih menantang karena gambarnya terlalu rumit. Kita harus mewakili gambar dengan tepat. Pada metode yang umum untuk mengatasinya adalah Bag of Visual Word (BOVW). Metode ini membutuhkan proses pengkodean untuk mendapatkan interpretasi data akhir. Ada banyak metode untuk melakukan pengkodean seperti Hard Quantization Coding (HQ), Sparse Coding (SC), dan Locality-constrained Linear Coding (LCC). Namun, metode pengkodean itu menggunakan asumsi yang berbeda. Oleh karena itu, kita harus membandingkan hasil setiap metode pengkodean. Metode pengkodean mempengaruhi akurasi klasifikasi. Metode pengkodean terbaik akan menghasilkan hasil klasifikasi yang lebih baik. Dataset UC Merced terdiri dari 21 kelas yang digunakan dalam penelitian ini. Hasil percobaan menunjukkan bahwa LCC memiliki kinerja / akurasi yang lebih baik daripada SC dan HQ. Metode LCC mendapat akurasi 86,48%. Selanjutnya, LCC juga mendapat performa terbaik pada berbagai jumlah data pelatihan untuk masing-masing kelas. Kata Kunci: Klasifikasi penggunaan lahan, citra penginderaan jauh beresolusi tinggi, Bag of Visual Word (BOVW), Sparse Coding (SC), Hard Quantization Coding (HQ)

show abstract

“…[1][2][3] One of the most basic tasks is to identify and map surface water boundaries. Optical remote sensing of water bodies is based on the difference in the spectral reflectance of land and water.…”

Section: Introductionmentioning

confidence: 99%

New hyperspectral difference water index for the extraction of urban water bodies by the use of airborne hyperspectral images

Xie

Luo

et al. 2014

J. Appl. Remote Sens

View full text Add to dashboard Cite

Abstract. Extracting surface land-cover types and analyzing changes are among the most common applications of remote sensing. One of the most basic tasks is to identify and map surface water boundaries. Spectral water indexes have been successfully used in the extraction of water bodies in multispectral images. However, directly applying a water index method to hyperspectral images disregards the abundant spectral information and involves difficulty in selecting appropriate spectral bands. It is also a challenge for a spectral water index to distinguish water from shadowed regions. The purpose of this study is therefore to develop an index that is suitable for water extraction by the use of hyperspectral images, and with the capability to mitigate the effects of shadow and low-albedo surfaces, especially in urban areas. Thus, we introduce a new hyperspectral difference water index (HDWI) to improve the water classification accuracy in areas that include shadow over water, shadow over other ground surfaces, and low-albedo ground surfaces. We tested the new method using PHI-2, HyMAP, and ROSIS hyperspectral images of Shanghai, Munich, and Pavia. The performance of the water index was compared with the normalized difference water index (NDWI) and the Mahalanobis distance classifier (MDC). With all three test images, the accuracy of HDWI was significantly higher than that of NDWI and MDC. Therefore, HDWI can be used for extracting water with a high degree of accuracy, especially in urban areas, where shadow caused by high buildings is an important source of classification error. © The Authors. Published by SPIE under a Creative Commons Attribution 3.0 Unported License. Distribution or reproduction of this work in whole or in part requires full attribution of the original publication, including its DOI.

show abstract

Improved VHR Urban Area Mapping Exploiting Object Boundaries

Cited by 67 publications

References 12 publications

Extraction of Built-Up Areas Using Convolutional Neural Networks and Transfer Learning From Sentinel-2 Satellite Images

Extraction of Built-Up Areas Using Convolutional Neural Networks and Transfer Learning From Sentinel-2 Satellite Images

Sparse Coding-Based Method Comparison For Land-Use Classification

New hyperspectral difference water index for the extraction of urban water bodies by the use of airborne hyperspectral images

Contact Info

Product

Resources

About