Research on UAV Image Road Recognition Application Based on Improved SRCNN
Authors:
- 6 December 2024
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 159-165
Abstract
The application of drones in surveying and monitoring is becoming increasingly important, especially in road information detection. Traditional road detection algorithms have low accuracy and poor anti-interference capabilities, while deep learning methods, although superior in performance, require a large amount of labeled data and have high computational complexity, making them difficult to apply in practice. To address this, this paper proposes an improved drone image road recognition method that combines SRCNN and UNet. The paper improves SRCNN by introducing an attention mechanism to enhance feature extraction capabilities and optimizes the network structure to reduce computational complexity. The improved SRCNN generates high-resolution images as input for UNet, which is further optimized to adapt to drone images. Experimental results show that this method outperforms existing technologies in terms of accuracy, real-time performance, and robustness, meeting practical requirements. In the future, we will explore combining other sensor data and optimizing network structures to enhance road detection performance.Keywords
This publication has 7 references indexed in Scilit:
- Remote Sensing Image Segmentation2024
- Analysis of weld pool region constituents in GMAW for dynamic reconstruction through characteristic enhancement and LSTM U-Net networksJournal of Manufacturing Processes, 2024
- Single-image super-resolution methods based on improved convolutional neural networks2024
- Ultrasound super resolution imaging for accurate uterus tumor detection and malignancy predictionJournal of Pharmaceutical and Biomedical Analysis Open, 2024
- Leveraging Data Analytics and a Deep Learning Framework for Advancements in Image Super-Resolution Techniques: From Classic Interpolation to Cutting-Edge Approaches2024
- NeighborLoss: A Loss Function Considering Spatial Correlation for Semantic Segmentation of Remote Sensing ImageIEEE Access, 2021
- DeepGlobe 2018: A Challenge to Parse the Earth through Satellite Images2018