TY - JOUR
T1 - A deep learning-based hyperspectral object classification approach via imbalanced training samples handling
AU - Islam, Md Touhid
AU - Islam, Md Rashedul
AU - Uddin, Md Palash
AU - Ulhaq, Anwaar
N1 - Publisher Copyright:
© 2023 by the authors.
PY - 2023/7
Y1 - 2023/7
N2 - Object classification in hyperspectral images involves accurately categorizing objects based on their spectral characteristics. However, the high dimensionality of hyperspectral data and class imbalance pose significant challenges to object classification performance. To address these challenges, we propose a framework that incorporates dimensionality reduction and re-sampling as preprocessing steps for a deep learning model. Our framework employs a novel subgroup-based dimensionality reduction technique to extract and select the most informative features with minimal redundancy. Additionally, the data are resampled to achieve class balance across all categories. The reduced and balanced data are then processed through a hybrid CNN model, which combines a 3D learning block and a 2D learning block to extract spectral–spatial features and achieve satisfactory classification accuracy. By adopting this hybrid approach, we simplify the model while improving performance in the presence of noise and limited sample size. We evaluated our proposed model on the Salinas scene, Pavia University, and Kennedy Space Center benchmark hyperspectral datasets, comparing it to state-of-the-art methods. Our object classification technique achieves highly promising results, with overall accuracies of 99.98%, 99.94%, and 99.46% on the three datasets, respectively. This proposed approach offers a compelling solution to overcome the challenges of high dimensionality and class imbalance in hyperspectral object classification.
AB - Object classification in hyperspectral images involves accurately categorizing objects based on their spectral characteristics. However, the high dimensionality of hyperspectral data and class imbalance pose significant challenges to object classification performance. To address these challenges, we propose a framework that incorporates dimensionality reduction and re-sampling as preprocessing steps for a deep learning model. Our framework employs a novel subgroup-based dimensionality reduction technique to extract and select the most informative features with minimal redundancy. Additionally, the data are resampled to achieve class balance across all categories. The reduced and balanced data are then processed through a hybrid CNN model, which combines a 3D learning block and a 2D learning block to extract spectral–spatial features and achieve satisfactory classification accuracy. By adopting this hybrid approach, we simplify the model while improving performance in the presence of noise and limited sample size. We evaluated our proposed model on the Salinas scene, Pavia University, and Kennedy Space Center benchmark hyperspectral datasets, comparing it to state-of-the-art methods. Our object classification technique achieves highly promising results, with overall accuracies of 99.98%, 99.94%, and 99.46% on the three datasets, respectively. This proposed approach offers a compelling solution to overcome the challenges of high dimensionality and class imbalance in hyperspectral object classification.
KW - batch normalization
KW - deep learning
KW - dimensionality reduction
KW - imbalanced data handling
KW - non-negative matrix factorization
KW - object classification
UR - http://www.scopus.com/inward/record.url?scp=85166181658&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85166181658&partnerID=8YFLogxK
UR - https://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=woscharlessturt_pure&SrcAuth=WosAPI&KeyUT=WOS:001036445600001&DestLinkType=FullRecord&DestApp=WOS
U2 - 10.3390/rs15143532
DO - 10.3390/rs15143532
M3 - Article
AN - SCOPUS:85166181658
SN - 2072-4292
VL - 15
SP - 1
EP - 29
JO - Remote Sensing
JF - Remote Sensing
IS - 14
M1 - 3532
ER -