드론 감시 시스템을 위한 인공지능 모델 학습 전략과 절대 좌표 추정 알고리즘
Artificial Intelligence Model Training Strategy and Absolute Coordinate Estimation Algorithm for Drone Surveillance Systems
1
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
최근 드론 기술의 혁신으로 인해 다양한 산업 분야에서의 응용 시도가 두드러지게 증가하고 있다. 이 중에서도, 기존의 고정형 CCTV 시스템을 대체하고자 드론을 활용한 감시 시스템을 구축하는 시도가 많아졌지만 센서의 무게 제한 및 캘리브레이션 문제와 같은 제약 사항이 발생한다. 이러한 어려움을 극복하고자, 본 논문에서는 인공지능 모델을 기반으로 한 드론 감시 시스템을 소개한다. 우리는 인공지능 모델의 성능을 향상시키기 위해 가상 환경에서 합성 데이터 수집 및 데이터 세트 구축을 수행하며, 실제 데이터 세트와 효과적으로 학습하는 몇 가지 학습 방법을 제시한다. 뿐만 아니라, 드론 환경에서 자주 발생하는 작은 객체를 효과적으로 탐지하기 위한 방법론을 적용하고, 탐지된 객체의 절대 좌표를 추정하기 위한 절대 좌표 추정 알고리즘을 제안한다. 더 나아가, 우리는 탐지된 객체에 대한 고유한 식별 번호(ID), 속도 측정, 인체의 관절 위치와 같은 추가적인 속성을 통해 드론 감시 시스템을 고도화하였으며, 이를 현실과 가상 환경에서 철저히 검증한다.
Abstract
Recently, drone technology innovations have led to a significant increase in application attempts in various industries. Among them, there are many attempts to build a surveillance system using drones to replace the existing fixed CCTV system, but constraints such as weight limitations and calibration problems of sensors occur. To overcome these difficulties, in this paper, we introduce a drone surveillance system based on an artificial intelligence(AI) model. To improve the performance of the AI model, we perform synthetic data collection and dataset construction in a virtual environment, and present several learning methods to effectively learn with real datasets. In addition, we apply a methodology to effectively detect small objects that often occur in the drone environment, and propose an absolute coordinate estimation algorithm to estimate the absolute coordinates of the detected objects. Furthermore, we enhance the drone surveillance system with additional attributes such as unique identification numbers (IDs), velocity measurements, and human keypoints for the detected objects, which are thoroughly validated in real and virtual environments.
참고문헌 (31)
- umme Juha, "A Study on Wireless LAN Attack Detection using Drones", THE JOURNAL OF KOREAN INSTITUTE OF NEXT GENERATION COMPUTING, 13(6), pp.31~43 (2017)
- Chanjeong Park, "Patent Trend Analysis of Anti-Drone : Focusing on the Neutralization Means and Methods", THE JOURNAL OF KOREAN INSTITUTE OF NEXT GENERATION COMPUTING, 16(2), pp.7~17 (2020)
- Hikmat Yar, "Efficient CNN Architecture for Forest Fire Detection Via Drones", 2021 THE JOURNAL OF KOREAN INSTITUTE OF NEXT GENERATION COMPUTING Spring Academic Conference, pp.341~345 (2021)
- Park Chae Hui, "Autonomous landing of Drone in No-GPS zone using vision sensor", THE JOURNAL OF KOREAN INSTITUTE OF NEXT GENERATION COMPUTING, 19(5), pp.32~44 (2023)
- SONG, Wenfeng, "Cross-view contextual relation transferred network for unsupervised vehicle tracking in drone videos", Proceedings of the IEEE, 2020, 1707~1716
- DOSOVITSKIY, Alexey, "CARLA: An open urban driving simulator", Conference on robot learning, pp.1~16 (2017), PMLR
- EVERINGHAM, Mark, "The pascal visual object classes(voc)challenge", International journal of computer vision, 88, pp.303~338 (2010)
- SUN, Pei, "Scalability in perception for autonomous driving: Waymo open dataset", Proceedings of the IEEE, 2020, 2446~2454
- LIN, Tsung-Yi, "Microsoft coco: Common objects in context", Computer Vision–ECCV 2014: 13th European Conference, 13, pp.740~755 (2014), Springer International Publishing
- SHAO, Shuai, "Objects365: A largescale, high-quality dataset for object detection", Proceedings of the IEEE, 2019, 8430~8439
- DU, Dawei, "The unmanned aerial vehicle benchmark: Object detection and tracking", Proceedings of the European conference on computer vision (ECCV), pp.370~386 (2018)
- WANG, Xiaolong, "A-fast-rcnn: Hard positive generation via adversary for object detection", Proceedings of the IEEE conference on computer vision and pattern recognition, pp.2606~2615 (2017)
- REN, Shaoqing, "Faster r-cnn: Towards real-time object detection with region proposal networks", Advances in neural information processing systems, pp.28~ (2015)
- WANG, Chien-Yao, "YOLOv7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors", Proceedings of the IEEE, 2023, 7464~7475
- BHAT, Shariq Farooq, "Localbins: Improving depth estimation by learning local distributions", European Conference on Computer Vision, pp.480~496 (2022), Springer Nature Switzerland
- MERTAN, Alican, "Single image depth estimation: An overview", Digital Signal Processing, 123, pp.103441~ (2022)
- RANFTL, René, "Towards robust monocular depth estimation : Mixing datasets for zero-shot cross-dataset transfer", IEEE transactions on pattern analysis and machine intelligence, 44(3), pp.1623~1637 (2020)
- ZHOU, Qianyu, "Context-aware mixup for domain adaptive semantic segmentation", IEEE Transactions on Circuits and Systems for Video Technology, 33(2), pp.804~817 (2022)
- CHEN, Lin, "Deliberated domain bridging for domain adaptive semantic segmentation", Advances in Neural Information Processing Systems, 35, pp.15105~15118 (2022)
- LI, Chun-Liang, "Cutpaste: Self-supervised learning for anomaly detection and localization", Proceedings of the IEEE, 2021, 9664~9674
- GOROSHIN, Ross, "Learning to linearize under uncertainty", Advances in neural information processing systems, pp.28~ (2015)
- SENSOY, Murat, "Evidential deep learning to quantify classification uncertainty", Advances in neural information processing systems, pp.31~ (2018)
- MADDOX, Wesley J., "A simple baseline for bayesian uncertainty in deep learning", Advances in neural information processing systems, pp.32~ (2019)
- CHEN, Ping-Yang, "Parallel residual bi-fusion feature pyramid network for accurate single-shot object detection", IEEE Transactions on Image Processing, 30, pp.9099~9111 (2021)
- XU, Yufei, "Vitpose: Simple vision transformer baselines for human pose estimation", Advances in Neural Information Processing Systems, 35, pp.38571~38584 (2022)
- GEIGER, Andreas, "Are we ready for autonomous driving? the kitti vision benchmark suite", 2012 IEEE conference on computer vision and pattern recognition, pp.3354~3361 (2012), IEEE