Deep learning based Machinery Threat Detection on Pipeline Right of Way
In this research, we develop a new deep learning strategy for robust detection and classification of objects on the pipeline right of way from aerial images. Our method can detect machinery threat with multiple sizes, different orientation and complex background in aerial images. In the proposed framework, the skip connection is used in the CNN structure to enhance feature learning. Also, the focal loss function is used to substitute for conventional cross entropy loss function in both of the region proposed network layer and the classifier layer.
K. Asari Vijayan
Primary Advisor's Department
Electrical and Computer Engineering
Stander Symposium poster
"Deep learning based Machinery Threat Detection on Pipeline Right of Way" (2019). Stander Symposium Posters. 1701.