İNSANSIZ HAVA ARAÇLARI İÇİN U-NET TABANLI ARAÇ TESPİT YÖNTEMİ

Abstract

Bilgisayar donanımı teknolojisindeki gelişmelerle birlikte bilgisayar görmesi ve yapay zeka alanlarındaki çalışmalar hız kazanmıştır. Bununla birlikte otonom sistemlerin kullanıldığı alanların sayısı da artmıştır. Bu alanlar arasında günümüz askeri teknolojisinin en önemli parametrelerinden biri olan insansız hava araçları yer almaktadır. İki farklı senaryoyu içeren bu çalışmamızda insansız hava araçlarının görüş yeteneklerini yapay zeka tabanlı olarak geliştirmeyi hedefledik. Senaryo-1 kapsamında ikili anlamsal bölütleme yöntemine uygun U-Net modeli sadece araç objesinin tespitini yapabilmek için insansız hava aracı kamerasıyla çekilen görüntüler yardımıyla eğitilmiştir. Hareketli veya durağan araç tespiti için tasarlanan Senaryo-2 kapsamında, U-Net modeli çok sınıflı anlamsal bölütlemeye uygun olarak eğitilmiştir. Tüm bu eğitim süreçlerinde kamuya açık veri seti kullanılmıştır. Senaryo-1 kapsamında eğitilen model %84,3 ortalama birleşim üzerinden kesişme (mIoU) değerine ulaşırken, Senaryo-2 kapsamında eğitilen model %79,7 mIoU değerine ulaşmıştır. Bu çalışmada yüksek çözünürlüklü görüntülerin model eğitiminde ve test aşamalarında kullanılabilmesi hakkında yaklaşımlar paylaşıldı. Bu tür çalışmaların sahada uygulanması, savunma sanayisinde hassaslığı ve güvenirliği iyileştirmeye yardımcı olabilir.

Keywords

• İHA
• U-Net
• Tespit
• Görüntü İşleme
• Anlamsal Bölütleme

U-NET BASED CAR DETECTION METHOD FOR UNMANNED AERIAL VEHICLES

Abstract

With the developments in computer hardware technology, studies in the fields of computer vision and artificial intelligence has accelerated. However, the number of areas where autonomous systems are used has also increased. Among these areas are unmanned aerial vehicles, which are one of the most important parameters of today's military technology. In this study, which includes two different scenarios, we aimed to improve the vision capabilities of unmanned aerial vehicles based on artificial intelligence. Within the scope of Scenario-1, the U-Net model suitable for binary semantic segmentation method was trained with the help of images taken by unmanned aerial vehicle camera. Within the scope of Scenario-2, which is designed for moving or stationary vehicle detection, the U-Net model is trained in accordance with multi-class semantic segmentation method. In all these training processes, a publicly available dataset was used. The model trained for Scenario-1 reached mean Intersection over Union (mIoU) value of 84.3%, while the model trained for Scenario-2 reached 79.7% mIoU. In this study, approaches were shared about the use of high-resolution images in model training and testing stages. Applying such studies in the field can help improve precision and reliability in arms industry.

Keywords

• UAV
• U-Net
• Detection
• Image Processing
• Semantic Segmentation

References

1. Nonami, K., Kendoul, F., Suzuki, S., Wang, W., Nakazawa, D., 2010. Autonomous flying robots: unmanned aerial vehicles and micro aerial vehicles. Springer Science & Business Media.
2. Boukoberine, M. N., Zhou, Z., Benbouzid, M., 2019. A critical review on unmanned aerial vehicles power supply and energy management: Solutions, strategies, and prospects. Applied Energy, 255, 113823.
3. Howard, J., Murashov, V., Branche, C. M., 2018. Unmanned aerial vehicles in construction and worker safety. American journal of industrial medicine, 61(1), 3-10.
4. Shareef, M. A., Kumar, V., Dwivedi, Y. K., Kumar, U., Akram, M. S., Raman, R., 2021. A new health care system enabled by machine intelligence: Elderly people's trust or losing self control. Technological Forecasting and Social Change, 162, 120334.
5. Mohamed, N., Al-Jaroodi, J., Jawhar, I., Idries, A., Mohammed, F., 2020. Unmanned aerial vehicles applications in future smart cities. Technological Forecasting and Social Change, 153, 119293.
6. Kuru, K., 2021. Planning the future of smart cities with swarms of fully autonomous unmanned aerial vehicles using a novel framework. IEEE Access, 9, 6571-6595.
7. Haulman, D. L., 2003. US unmanned aerial vehicles in combat, 1991-2003. AIR FORCE HISTORICAL RESEARCH AGENCY MAXWELL AFB AL.
8. Xu, Y., Yu, G., Wu, X., Wang, Y., Ma, Y., 2016. An enhanced Viola-Jones vehicle detection method from unmanned aerial vehicles imagery. IEEE Transactions on Intelligent Transportation Systems, 18(7), 1845-1856.

9. Dargan, S., Kumar, M., Ayyagari, M. R., Kumar, G., 2020. A survey of deep learning and its applications: a new paradigm to machine learning. Archives of Computational Methods in Engineering, 27(4), 1071-1092.
10. Blumberg, S. B., Tanno, R., Kokkinos, I., Alexander, D. C., 2018. Deeper image quality transfer: Training low-memory neural networks for 3d images. International Conference on Medical Image Computing and Computer-Assisted Intervention, 118-125.
11. Du, D., Qi, Y., Yu, H., Yang, Y., Duan, K., Li, G., Tian, Q., 2018. The unmanned aerial vehicle benchmark: Object detection and tracking. European conference on computer vision, 370-386.
12. Mueller, M., Smith, N., Ghanem, B., 2016. A benchmark and simulator for uav tracking. European conference on computer vision, 445-461.
13. Zhao, T., Nevatia, R., 2003. Car detection in low resolution aerial images. Image and vision computing, 21(8), 693-703.
14. Ammour, N., Alhichri, H., Bazi, Y., Benjdira, B., Alajlan, N., Zuair, M., 2017. Deep learning approach for car detection in UAV imagery. Remote Sensing, 9(4), 312.
15. Xu, Y., Yu, G., Wang, Y., Wu, X., Ma, Y., 2017. Car detection from low-altitude UAV imagery with the faster R-CNN. Journal of Advanced Transportation.
16. Hinz, S., Stilla, U., 2006. Car detection in aerial thermal images by local and global evidence accumulation. Pattern Recognition Letters, 27(4), 308-315.
17. Lyu, Y., Vosselman, G., Xia, G. S., Yilmaz, A., Yang, M. Y., 2020. UAVid: A semantic segmentation dataset for UAV imagery. ISPRS Journal of Photogrammetry and Remote Sensing, 165, 108-119.