Technology Category
Computer Vision: Remote Sensing / Geospatial AIMachine Learning: Multi-class/Multi-label Learning & Extreme ClassificationIntelligent Robots: Multimodal Perception & Sensor Fusion
Remote Sensing Image Classification Using Convolutional Neural Network (CNN) and Transfer Learning Techniques
🤖 AI Summary
This work addresses the classification of aerial remote sensing images depicting four complex land-cover classes: transmission towers, forests, farmlands, and mountainous terrain. To this end, we construct a dedicated dataset comprising 10,400 annotated images—fused from self-collected Google satellite imagery and the MLRNet dataset. Building upon a CNN-Softmax framework, we systematically benchmark transfer learning performance using VGG16 and MobileNetV2, and further train a custom CNN baseline. Our study is the first to empirically demonstrate MobileNetV2’s superiority for lightweight remote sensing scene classification: it achieves 96% test accuracy (loss = 0.119), significantly outperforming VGG16 (90%, loss = 0.298) and the custom CNN (87%), while maintaining high inference efficiency and robustness under limited training samples. This represents the state-of-the-art performance for this classification task and provides an efficient, deployable solution for edge applications such as power line inspection.
Technology Category
Application Category
📝 Abstract
This study investigates the classification of aerial images depicting transmission towers, forests, farmland, and mountains. To complete the classification job, features are extracted from input photos using a Convolutional Neural Network (CNN) architecture. Then, the images are classified using Softmax. To test the model, we ran it for ten epochs using a batch size of 90, the Adam optimizer, and a learning rate of 0.001. Both training and assessment are conducted using a dataset that blends self-collected pictures from Google satellite imagery with the MLRNet dataset. The comprehensive dataset comprises 10,400 images. Our study shows that transfer learning models and MobileNetV2 in particular, work well for landscape categorization. These models are good options for practical use because they strike a good mix between precision and efficiency; our approach achieves results with an overall accuracy of 87% on the built CNN model. Furthermore, we reach even higher accuracies by utilizing the pretrained VGG16 and MobileNetV2 models as a starting point for transfer learning. Specifically, VGG16 achieves an accuracy of 90% and a test loss of 0.298, while MobileNetV2 outperforms both models with an accuracy of 96% and a test loss of 0.119; the results demonstrate the effectiveness of employing transfer learning with MobileNetV2 for classifying transmission towers, forests, farmland, and mountains.
Problem
Research questions and friction points this paper is trying to address.
Classifying aerial images using CNN and transfer learning.
Evaluating MobileNetV2 and VGG16 for landscape categorization.
Achieving high accuracy in image classification tasks.
Innovation
Methods, ideas, or system contributions that make the work stand out.
Uses CNN and Softmax for image classification.
Applies transfer learning with MobileNetV2 and VGG16.
Achieves high accuracy with MobileNetV2 (96%).
🔎 Similar Papers
No similar papers found.
