More than Just Lines on a Map: Best Practices for U.S Bike Routes
Improve Divers Tracking and Classification in Sonar Images Using Robust Diver Wake Detection Algorithm
1. Improve Divers Tracking and Classification in Sonar Images
Using Robust Diver Wake Detection Algorithm
Mohammad Tarek Al Muallim
Ozhan Duzenli
Ceyhun Ilguy
ICMLC 2017 : 19th International Conference on Machine Learning and Computing
Vancouver, Canada
2. Problem Description and Challenges
▪ Diver detection active sonar.
▪ Detect underwater threats such as divers and AUV.
▪ Deal with very noisy sonar images.
▪ Detect, Track and Classify of underwater objects.
2
3. Method
▪ Three stages : detect, track and classify.
▪ Detect:
▪ sonar images is normalized.
▪ segmented based on fixed threshold.
▪ centroids of the segments are found and clustered based on distance metric.
▪ Track:
▪ Linear Kalman filter is applied and fine tuned.
▪ Filtering stage to eliminate the noisy and unstable tracks.
▪ Eliminate object with a speed out of the diver speed range.
3
4. Method
▪ Classify:
▪ Deiced wither if the tracked object is an open circuit diver or a close circuit diver.
▪ Small area around the object is extracted.
▪ Novel wake detection method is applied.
▪ Morphological features of the object with his wake is extracted.
▪ Support vector machine used to find the best classifier.
4
5. Results
▪ The sonar training images and the test images are collected by ARMELSAN Defense
Technologies Company using the portable diver detection sonar ARAS-2023.
▪ After applying the algorithm to the test sonar data, we get fine and stable tracks of
the divers.
▪ The total classification accuracy achieved with the diver type is 97%.
5