Airplane Detection Based on Mask Region Convolution Neural Network
This provides a more accessible solution for airport traffic management in resource-limited settings, though it is incremental as it adapts existing deep learning methods to a new data source.
The paper tackled airplane detection at airports by proposing a low-cost drone-based system using a modified Mask RCNN model, achieving promising results on COCO metrics.
Addressing airport traffic jams is one of the most crucial and challenging tasks in the remote sensing field, especially for the busiest airports. Several solutions have been employed to address this problem depending on the airplane detection process. The most effective solutions are through the use of satellite images with deep learning techniques. Such solutions, however, are significantly costly and require satellites and modern complicated technology which may not be available in most countries worldwide. This paper provides a universal, low cost and fast solution for airplane detection in airports. This paper recommends the use of drones instead of satellites to feed the system with drone images using a proposed deep learning model. Drone images are employed as the dataset to train and evaluate a mask region convolution neural network (RCNN) model. The Mask RCNN model applies faster RCNN as its base configuration with critical modifications on its head neural network constructions. The model detects whether or not an airplane is present and includes mask estimations to approximate surface area and length, which will help future works identify the airplane type. This solution can be easily implemented globally as it is a low-cost and fast solution for airplane detection at airports. The evaluation process reveals promising results according to Microsoft Common Objects in Context (COCO) metrics.