Search results
1 – 2 of 2
The purpose of this paper is to present a deep ensemble neural network model for the detection of forest fires in aerial vehicle videos.
Abstract
Purpose
The purpose of this paper is to present a deep ensemble neural network model for the detection of forest fires in aerial vehicle videos.
Design/methodology/approach
Presented deep ensemble models include four convolutional neural networks (CNNs): a faster region-based CNN (Faster R-CNN), a simple one-stage object detector (RetinaNet) and two different versions of the you only look once (Yolo) models. The presented method generates its output by fusing the outputs of these different deep learning (DL) models.
Findings
The presented fusing approach significantly improves the detection accuracy of fire incidents in the input data.
Research limitations/implications
The computational complexity of the proposed method which is based on combining four different DL models is relatively higher than that of using each of these models individually. On the other hand, however, the performance of the proposed approach is considerably higher than that of any of the four DL models.
Practical implications
The simulation results show that using an ensemble model is quite useful for the precise detection of forest fires in real time through aerial vehicle videos or images.
Social implications
By this method, forest fires can be detected more efficiently and precisely. Because forests are crucial breathing resources of the earth and a shelter for many living creatures, the social impact of the method can be considered to be very high.
Originality/value
This study fuses the outputs of different DL models into an ensemble model. Hence, the ensemble model provides more potent and beneficial results than any of the single models.
Details
Keywords
Nurcan Sarikaya Basturk and Abdurrahman Sahinkaya
The purpose of this paper is to present a detailed performance comparison of recent and state-of-the-art population-based optimization algorithms for the air traffic control…
Abstract
Purpose
The purpose of this paper is to present a detailed performance comparison of recent and state-of-the-art population-based optimization algorithms for the air traffic control problem.
Design/methodology/approach
Landing sequence and corresponding landing times for the aircrafts were determined by using population-based optimization algorithms such as artificial bee colony, particle swarm, differential evolution, biogeography-based optimization, simulated annealing, firefly and teaching–learning-based optimization. To obtain a fair comparison, all simulations were repeated 30 times for each of the seven algorithms, two different problems and two different population sizes, and many different criteria were used.
Findings
Compared to conventional methods that depend on a single solution at the same time, population-based algorithms have simultaneously produced many alternate possible solutions that can be used recursively to achieve better results.
Research limitations/implications
In some cases, it may take slightly longer to obtain the optimum landing sequence and times compared to the methods that give a direct result; however, the processing times can be reduced using powerful computers or GPU computations.
Practical implications
The simulation results showed that using population-based optimization algorithms were useful to obtain optimal landing sequence and corresponding landing times. Thus, the proposed air traffic control method can also be used effectively in real airport applications.
Social implications
By using population-based algorithms, air traffic control can be performed more effectively. In this way, there will be more efficient planning of passengers’ travel schedules and efficient airport operations.
Originality/value
The study compares the performances of recent and state-of-the-art optimization algorithms in terms of effective air traffic control and provides a useful approach.
Details