Land-Use Classification via Extreme Learning Classifier Based on Deep Convolutional Features

• Published in: IEEE geoscience and remote sensing letters 2017-05, Vol.14 (5), p.704-708
• Main Authors: Weng, Qian, Mao, Zhengyuan, Lin, Jiawen, Guo, Wenzhong
• Format: Article
• Language: English
• Subjects: Accuracy; Artificial neural networks; Classification; Computer architecture; Convolutional neural network (CNN); Deep learning; Effectiveness; extreme learning machine (ELM); Feature extraction; Image quality; Image resolution; Land use; Land use classification; Learning algorithms; Machine learning; Methods; Neural networks; Neurons; Remote sensing; scene understanding; Support vector machines; Training
• ISSN: 1545-598X; 1558-0571
• DOI: 10.1109/LGRS.2017.2672643

One of the challenging issues in high-resolution remote sensing images is classifying land-use scenes with high quality and accuracy. An effective feature extractor and classifier can boost classification accuracy in scene classification. This letter proposes a deep-learning-based classification method, which combines convolutional neural networks (CNNs) and extreme learning machine (ELM) to improve classification performance. A pretrained CNN is initially used to learn deep and robust features. However, the generalization ability is finite and suboptimal, because the traditional CNN adopts fully connected layers as classifier. We use an ELM classifier with the CNN-learned features instead of the fully connected layers of CNN to obtain excellent results. The effectiveness of the proposed method is tested on the UC-Merced data set that has 2100 remotely sensed land-use-scene images with 21 categories. Experimental results show that the proposed CNN-ELM classification method achieves satisfactory results.