Airplane Type Identification Based on Mask RCNN; An Approach to Reduce Airport Traffic Congestion

Al-Shaibani, W.T; Helvaci, Mustafa; Shayea, Ibraheem; Azizan, Azızul

doi:10.1109/mttw53539.2021.9607203

Cited by 7 publications

(1 citation statement)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Deep Convolutional Neural Network (DCNN) with the use of Principal Component Analysis (PCA) for feature selection [19] and detection and recognition with Single Shot multi-box Detector (SSD) network [20]. Other approaches, such as bilinear discriminative Extreme Learning Machine (ELM) network (BD-ELMNet) [21] is used for aircraft recognition, and Region-based Convolutional Neural Networks (RCNN) [22] is used for aircraft identification, which is done on remote sensing images.…”

Section: ) Aircraft Recognition or Classification Using Neuralmentioning

confidence: 99%

Intelligent Military Aircraft Recognition and Identification to Support Military Personnel on the Air Observation Operation

Sumari¹,

Adinandra²,

Syulistyo³

et al. 2022

Int. J. Adv. Sci. Eng. Inf. Technol.

View full text Add to dashboard Cite

A hostile or unfriendly aircraft will mostly fly at low-level altitude or hide behind natural obstacles to avoid Radar detection. One of the ways to detect and recognize while at the same time identifying such aircraft is to perform air observation from the ground. A technique called Visual Aircraft Recognition (VACR) has been practiced in training soldiers to recognize and find an incoming aircraft from a distance using binoculars. Remembering so many types of aircraft have their challenge. To ease the task, we have designed and developed an intelligent military aircraft recognition and identification system using the combination of Back Propagation Neural Networks (BPNN) and Information Fusion to speed up the recognition and identification. We use 13 aircraft features fused into five primary ones as the inputs to the BPNN for the recognition, while the identification uses Hamming Distance to the recognition results. With 155 data consisting of 85 military aircraft and helicopters and 70 civilian aircraft and helicopters and applying the 80:20 scheme for the training and test data, our system can obtain 95.33% and 87% accuracy at the training phase and the test phase. It also succeeds in recognizing and identifying a new military aircraft that is not in the dataset, while the Information Fusion can speed up the recognition and identification by up to 6 seconds. This impacts the acceleration of aircraft recognition and identification.

show abstract

Section: ) Aircraft Recognition or Classification Using Neuralmentioning

confidence: 99%