Improved Metric Learning With the CNN for Very-High-Resolution Remote Sensing Image Classification

Cheng, Shijie; Lv, Zhiyong; Shen, Huifang; Fang, Li; You, Zhenzhen

doi:10.1109/jstars.2020.3033944

Cited by 8 publications

(5 citation statements)

References 43 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The weighted CNN-based spatial features are then fed into the RF classifier. Another possible approach is a metric learning network in which the similarity between high-dimensional features is evaluated during training the CNN model [61]. Thus, it is worthwhile to investigate the potential of the improved feature learning scheme in future work.…”

Section: Discussionmentioning

confidence: 99%

Potential of Hybrid CNN-RF Model for Early Crop Mapping with Limited Input Data

et al. 2021

View full text Add to dashboard Cite

When sufficient time-series images and training data are unavailable for crop classification, features extracted from convolutional neural network (CNN)-based representative learning may not provide useful information to discriminate crops with similar spectral characteristics, leading to poor classification accuracy. In particular, limited input data are the main obstacles to obtain reliable classification results for early crop mapping. This study investigates the potential of a hybrid classification approach, i.e., CNN-random forest (CNN-RF), in the context of early crop mapping, that combines the automatic feature extraction capability of CNN with the superior discrimination capability of an RF classifier. Two experiments on incremental crop classification with unmanned aerial vehicle images were conducted to compare the performance of CNN-RF with that of CNN and RF with respect to the length of the time-series and training data sizes. When sufficient time-series images and training data were used for the classification, the accuracy of CNN-RF was slightly higher or comparable with that of CNN. In contrast, when fewer images and the smallest training data were used at the early crop growth stage, CNN-RF was substantially beneficial and the overall accuracy increased by maximum 6.7%p and 4.6%p in the two study areas, respectively, compared to CNN. This is attributed to its ability to discriminate crops from features with insufficient information using a more sophisticated classifier. The experimental results demonstrate that CNN-RF is an effective classifier for early crop mapping when only limited input images and training samples are available.

show abstract

Section: Discussionmentioning

confidence: 99%

Potential of Hybrid CNN-RF Model for Early Crop Mapping with Limited Input Data

et al. 2021

View full text Add to dashboard Cite

show abstract

“…The proposed methodology was compared on two datasets and the highest overall accuracy of 98.44% was obtained on the Houston University dataset. Shi et al, 142 proposed a semisupervised method for the classification of hyperspectral images through improved metric learning (IML) and convolutional neural network known as IML–CNN and used pixel constraint for selecting unlabeled samples. Based on high dimensional features that are extracted by 3D CNN, IML is constructed to predict the distance between each unlabeled sample and each class center.…”

Section: Methodsmentioning

confidence: 99%

Support vector machine versus convolutional neural network for hyperspectral image classification: A systematic review

Kaul

2022

Concurrency and Computation

View full text Add to dashboard Cite

Various machine learning and deep learning techniques have been proposed for classification purposes in the case of hyperspectral imaging. Among all the machine learning techniques support vector machine (SVM) has been a promising classification algorithm in the case of remote sensing applications, particularly in the field of hyperspectral imaging. In the case of deep learning techniques, the convolutional neural network is gaining much attention from researchers for classification purposes and has given reliable results in this field, thus is being used widely by researchers. So, in this article, a systematic review about the two most used hyperspectral classification techniques that is, SVM and CNN is given. A total of 86 papers from the four well-known journals belonging to the field of remote sensing have been reviewed where the methodologies used by different authors, datasets used, results acquired, contribution and shortcomings have been put forth. This meta-analysis generally focuses on the recent SVM and CNN methodologies that are proposed by various authors. A summary of the best results obtained from SVM and CNN in terms of classification accuracies is also provided to help the researchers working in this area achieve better results.

show abstract

“…Since CNNs achieve excellent results in the ImageNet Large Scale Visual Recognition Challenge 2012 (ILSVRC2012) [82], they have been widely used in remote sensing image classification [83][84][85][86][87]. The CNN method achieves higher accuracy in the extraction of urban impervious surfaces than the traditional method [81].…”

Section: Deep Learning Modelsmentioning

confidence: 99%

Emerging Issues in Mapping Urban Impervious Surfaces Using High-Resolution Remote Sensing Images

Shao

Cheng

et al. 2023

Remote Sensing

View full text Add to dashboard Cite

Urban impervious surface (UIS) is a key parameter in climate change, environmental change, and sustainability. UIS extraction has been evolving rapidly in the past decades. However, high-resolution impervious surface mapping is a long-term need. There is an urgent requirement for impervious surface mapping from high-resolution remote sensing imagery. In this paper, we compare current extraction methods in terms of extraction units and extraction models and summarize their strengths and limitations. We discuss the challenges in impervious surface estimation from high spatial resolution remote sensing imagery in terms of selection of spatial resolution, spectral band, and extraction method. The uncertainties caused by clouds and snow, shadows, and vegetation occlusion are also analyzed. Automated sample labeling and remote sensing domain knowledge are the main directions in impervious surface extraction using deep learning methods. We should also focus on using continuous time series of high-resolution imagery and multi-source satellite imagery for dynamic monitoring of impervious surfaces.

show abstract

Improved Metric Learning With the CNN for Very-High-Resolution Remote Sensing Image Classification

Cited by 8 publications

References 43 publications

Potential of Hybrid CNN-RF Model for Early Crop Mapping with Limited Input Data

Potential of Hybrid CNN-RF Model for Early Crop Mapping with Limited Input Data

Support vector machine versus convolutional neural network for hyperspectral image classification: A systematic review

Emerging Issues in Mapping Urban Impervious Surfaces Using High-Resolution Remote Sensing Images

Contact Info

Product

Resources

About