Batch-Mode Active Learning-Based Superpixel Library Generation for Very High-Resolution Aerial Image Classification

2018 ◽  
pp. 307-318
Author(s):  
Rajeswari Balasubramaniam ◽  
Srivalsan Namboodiri ◽  
Gorthi. R. K. Sai Subrahmanyam ◽  
Rama Rao Nidamanuri
Keyword(s):  
High Resolution ◽  
Active Learning ◽  
Image Classification ◽  
Aerial Image ◽  
Batch Mode ◽  
Very High

Batch Mode Active Learning for Geographical Image Classification

2015 ◽  
pp. 744-755
Author(s):  
Zengmao Wang ◽  
Bo Du ◽  
Lefei Zhang ◽  
Wenbin Hu ◽  
Dacheng Tao ◽  
...  
Keyword(s):  
Active Learning ◽  
Image Classification ◽  
Batch Mode

scholarly journals A ROUGH SET DECISION TREE BASED MLP-CNN FOR VERY HIGH RESOLUTION REMOTELY SENSED IMAGE CLASSIFICATION

2017 ◽  
Vol XLII-2/W7 ◽  
pp. 1451-1454 ◽  
Author(s):  
C. Zhang ◽  
X. Pan ◽  
S. Q. Zhang ◽  
H. P. Li ◽  
P. M. Atkinson
Keyword(s):  
High Resolution ◽  
Decision Tree ◽  
Image Classification ◽  
Rough Set ◽  
Remotely Sensed ◽  
Classification Performance ◽  
Learning Approaches ◽  
Remotely Sensed Data ◽  
Remotely Sensed Image ◽  
Very High

Recent advances in remote sensing have witnessed a great amount of very high resolution (VHR) images acquired at sub-metre spatial resolution. These VHR remotely sensed data has post enormous challenges in processing, analysing and classifying them effectively due to the high spatial complexity and heterogeneity. Although many computer-aid classification methods that based on machine learning approaches have been developed over the past decades, most of them are developed toward pixel level spectral differentiation, e.g. Multi-Layer Perceptron (MLP), which are unable to exploit abundant spatial details within VHR images. <br><br> This paper introduced a rough set model as a general framework to objectively characterize the uncertainty in CNN classification results, and further partition them into correctness and incorrectness on the map. The correct classification regions of CNN were trusted and maintained, whereas the misclassification areas were reclassified using a decision tree with both CNN and MLP. The effectiveness of the proposed rough set decision tree based MLP-CNN was tested using an urban area at Bournemouth, United Kingdom. The MLP-CNN, well capturing the complementarity between CNN and MLP through the rough set based decision tree, achieved the best classification performance both visually and numerically. Therefore, this research paves the way to achieve fully automatic and effective VHR image classification.

scholarly journals Very-high-resolution mapping of river-immersed topography by remote sensing

2008 ◽  
Vol 32 (4) ◽  
pp. 403-419 ◽  
Author(s):  
Denis Feurer ◽  
Jean-Stéphane Bailly ◽  
Christian Puech ◽  
Yann Le Coarer ◽  
Alain A. Viau
Keyword(s):  
Remote Sensing ◽  
High Resolution ◽  
Large Scale ◽  
Aerial Image ◽  
Depth Measurement ◽  
Very High Spatial Resolution ◽  
Remote Sensing Techniques ◽  
Logarithmic Relationship ◽  
Ground Penetrating ◽  
Very High

Remote sensing has been used to map river bathymetry for several decades. Non-contact methods are necessary in several cases: inaccessible rivers, large-scale depth mapping, very shallow rivers. The remote sensing techniques used for river bathymetry are reviewed. Frequently, these techniques have been developed for marine environment and have then been transposed to riverine environments. These techniques can be divided into two types: active remote sensing, such as ground penetrating radar and bathymetric lidar; or passive remote sensing, such as through-water photogrammetry and radiometric models. This last technique — which consists of finding a logarithmic relationship between river depth and image values — appears to be the most used. Fewer references exist for the other techniques, but lidar is an emerging technique. For each depth measurement method, we detail the physical principles and then a review of the results obtained in the field. This review shows a lack of data for very shallow rivers, where a very high spatial resolution is needed. Moreover, the cost related to aerial image acquisition is often huge. Hence we propose an application of two techniques, radiometric models and through-water photogrammetry, with very- high-resolution passive optical imagery, light platforms, and off-the-shelf cameras. We show that, in the case of the radiometric models, measurement is possible with a spatial filtering of about 1 m and a homogeneous river bottom. In contrast, with through-water photogrammetry, fine ground resolution and bottom textures are necessary.

Sign in / Sign up

Export Citation Format

Share Document