자율협력주행을 위한 센서전용시설물의 데이터셋 구축
Building Dataset of Sensor-only Facilities for Autonomous Cooperative Driving
0
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
본 논문에서는 자율협력주행 인프라를 위해 제작된 8가지 센서 전용 시설물들에 대해 라이다로 취득한 포인트 클라우드 데이터로부터 시설물들의 특징을 추출하여 샘플 데이터셋으로 구축하는 방법을 제안한다. 고휘도 반사지가 부착된 8가지 센서 전용 시설물들과 데이터 취득 시스템을 개발했고, 취득된 포인트 클라우드 데이터로부터 일정한 측정 거리 내에 위치한 시설물들의 특징을 추출하기위해 포인트 대상의 DBSCAN 방법과 반사강도 대상의 OTSU 방법을 응용하여 추려낸 포인트들에 원통형 투영법을 적용했다. 3차원 포인트 좌표, 2차원 투영 좌표, 그리고 반사강도 등을 해당 시설물의특징으로 설정했고, 정답 레이블과 함께 데이터셋으로 제작했다. 라이다로 취득한 데이터를 기반으로구축된 시설물 데이터셋의 효용 가능성을 확인하기 위해서 기본적인 CNN 모델을 선정하여 학습 후테스트를 진행하여 대략 90% 이상의 정확도를 보여 시설물 인식 가능성을 확인했다. 지속적인 실험을통해 제시한 데이터셋 구축을 위한 특징 추출 알고리즘의 개선 및 성능 향상과 더불어 이에 적합한자율협력주행을 위한 센서 전용 시설물을 인식할 수 있는 전용 모델을 개발할 예정이다.
Abstract
In this paper, we propose a method to build a sample dataset of the features of eight sensor-only facilities built as infrastructure for autonomous cooperative driving. The feature extracted from point cloud data acquired by LiDAR and build them into the sample dataset for recognizing the facilities. In order to build the dataset, eight sensor-only facilities with high-brightness reflector sheets and a sensor acquisition system were developed. To extract the features of facilities located within a certain measurement distance from the acquired point cloud data, a cylindrical projection method was applied to the extracted points after applying DBSCAN method for points and then a modified OTSU method for reflected intensity. Coordinates of 3D points, projected coordinates of 2D, and reflection intensity were set as the features of the facility, and the dataset was built along with labels. In order to check the effectiveness of the facility dataset built based on LiDAR data, a common CNN model was selected and tested after training, showing an accuracy of about 90% or more, confirming the possibility of facility recognition. Through continuous experiments, we will improve the feature extraction algorithm for building the proposed dataset and improve its performance, and develop a dedicated model for recognizing sensor-only facilities for autonomous cooperative driving.
참고문헌 (16)
- Sang-Yeon Won, "A Comparison of Korea Standard HD Map for Actual Driving Support of Autonomous Vehicles and Analysis of Application Layers", Journal of the Korean Association of Geographic Information Studies, 23(3), pp.132~145 (2020)[DOI]
- Young-Jae Jeon, "Study on the Material Characteristics of Improved and Virtual Facility for Autonomous Vehicles", Journal of Korea Contetns Association, 22(12), pp.171~179 (2022)[DOI]
- Young-Jae Jeon, "Deriving the Role of Sign Facilities Recognized by Autonomous Vehicles", Journal of the Korean Association of Geographic Information Studies, 26(1), pp.1~10 (2023)[DOI]
- Young-Jae Jeon, "Study on Applying New Infrastructure for Autonomous Driving in HD Maps", Journal of the Korea Association of Geographic Information Studies, 26(4), pp.1~11 (2023)[DOI]
- Ji-Yeon Kim, "A Research of Factors Affecting LiDAR’s Detection on Road Signs : Focus on Shape and Height of Road Sign", Journal of Korea Institute of Intelligence Transport Systems, 21(4), pp.190~211 (2022)[DOI]
- Bum-Jin Park, "A Study of LiDAR’s Performance Change by Road Sign’s Color and Climate", Journal of Korea Institute of Intelligence Transport Systems, 20(6), pp.227~241 (2021)[DOI]
- Do Young Kim, "A Study on data set construction method using deep learning technologies –improvement of accuracy through dataset and hyper-parameter adjustment using CNN", Journal of Digital Contents Society, 24(2), pp.341~351 (2023)[DOI]
- "Building Guide of AI Training Dataset", Ministry of Science and ICT, 157~220 (2021)
- Charles R. Qi, "PointNet : Deep Learning on Point Sets for 3D Classification and Segmentation", Proceedings on the IEEE Conference on Computer Vision and Pattern Recognition(CVPR), pp.252~660 (2017)
- Charles R. Qi, "PointNet++ : Deep Hierarchical Feature Learning on Point Sets in a Metric Space", Proceedings of the 30th Conference on Neural Infroamtion Processing System(NeurIPS 2017) (2017)
- Yangyn Li, "PointCNN : Convolution On x-Trans formed Points", Proceedings of the 32nd Conference on Neural Informatio Processing Systems(NeurIPS 2018), pp. 828-838, 2018, pp.828~838 (2018)
- Aoran Xiao, "FPS-Net : A convolutional fusion network for large-scale LiDAR point cloud segmentation", ISPRS Journal of Photogrammetry and Remote Sensing, 17, pp.237~249 (2021)[DOI]
- Jaehyun Park, "PCSCNet : Fast 3D semantic segmentation of LiDAR point cloud for autonomous car using point convolution and sparse convolution network", Export Systems with Application, 212 (2023)[DOI]
- Xiakhan Tu, "LiDAR Point Cloud Recognition and Visualization with Deep Learning for Overhead Contact Inspection", Sensors, 20, pp.6387~ (2020)[DOI]
- J. Sreevalsan-Nair, "Geospatial Infrastructure, Applications and Technologies: India Case Studies", pp.313~325, Springer
- Alexey Dosovotskiy, "An Image is Worth 16x16 Words:Transformers for Image Recognition at Scale", Published as a conference paper at ICLR 2021 (2021)