• The Journal of the Korea institute of electronic communication sciences
• /
• v.13 no.2
• /
• pp.313-320
• /
• 2018

The importance of sensors on a self-driving vehicle has rising since it act as eyes for the vehicle. Lidar sensors based on laser technology tend to yield better image quality with more laser channels, thus, it has higher detection accuracy for obstacles, pedistrians, terrain, and other vechicles. However, incorporating more laser channels results higher unit price more than ten times, and this is a major drawback for using high channel lidar sensors on a vehicle for actual consumer market. To come up with this drawback, we propose a method of integrating multiple low channel, low cost lidar sensors acting as one high channel sensor. The result uses four 16 channels lidar sensors with small form factor acting as one bulky 64 channels sensor, which in turn, improves vehicles cosmetic aspects and helps widespread of using the lidar technology for the market.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2016.06a
• /
• pp.200-203
• /
• 2016

최근 자율주행자동차에 대한 관심이 증가함에 따라 교통 상황을 인식하는 방법에 대한 연구도 활발하게 진행되고 있다. 특히 교통신호등의 인식은 치명적인 결과를 야기하는 교통사고와 밀접하게 연관된다는 점에서 중요성이 더욱 부각되고 있다. 본 논문에서는 컴퓨터 비전 시스템을 기반으로 한 교통신호등 인식 방법을 제안한다. 차선, 표지판 등과는 다르게 교통신호등은 빛을 발하는 특징이 있으며 그 모양과 형태 또한 규격화 되어 있다. 이러한 특징 중 색상과 형태 특징을 이용하여 두 단계의 추출과정을 거쳐 교통신호등을 인식한다. 먼저 HSV 색 공간에서 적색, 녹색, 주황색의 빛을 발하는 영역을 찾아낸 뒤, 신호의 원형 특징을 이용해 가로, 세로 사이즈와 크기로 신호의 후보를 추출한다. 다음, 신호등의 검은 박스 영역을 찾기 위해 추출한 신호 후보군의 주변부가 검정색인지를 확인한다. 최종적으로 신호등의 박스 부분을 검출하여 신호를 발하는 위치를 기반으로 신호를 인식한다. 실험결과 많은 계산량을 요구하는 기계학습을 사용하지 않고도 실시간 처리와 높은 인식률로 교통 신호를 인식할 수 있음을 확인하였다.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.19 no.5
• /
• pp.19-24
• /
• 2019

The development of ICT technology has a huge impact on the existing closed CCTV security equipment market. With the importance of video data particularly highlighted in areas such as self-driving cars, unmanned vehicles and smart cities, various technologies using video are emerging. In this paper, we proposed a method to transmit videos and metadata as a part of smart city integration, and to solve the traffic, environment and security problems caused in urban life by utilizing the metadata instead of using CCTV videos for simple recording purposes, and reverse tunneling technique was designed and implemented as a method for accessing CCTV videos for vehicles from remote locations. Integrated management of CCTV videos and metadata for vehicles that have been used only for limited purposes in closed environments will enable efficient operation of integrated centers in real time required by smart cities, such as vehicle status check, road conditions and facility management.

• Journal of Broadcast Engineering
• /
• v.26 no.6
• /
• pp.738-747
• /
• 2021

Today, with the improvement of deep learning technology, computer vision areas such as image classification, object detection, object segmentation, and object tracking have shown remarkable improvements. Various applications such as intelligent surveillance, robots, Internet of Things, and autonomous vehicles in combination with deep learning technology are being applied to actual industries. Accordingly, the requirement of an efficient compression method for video data is necessary for machine consumption as well as for human consumption. In this paper, we propose an object-based compression of thermal infrared images for machine vision. The input image is divided into object and background parts based on the object detection results to achieve efficient image compression and high neural network performance. The separated images are encoded in different compression ratios. The experimental result shows that the proposed method has superior compression efficiency with a maximum BD-rate value of -19.83% to the whole image compression done with VVC.

• Journal of The Korean Association of Information Education
• /
• v.25 no.6
• /
• pp.947-960
• /
• 2021

In this study, we examine the AI ethics perception of university students to explore the direction of AI ethics education. For this, 83 students wrote their thoughts about 5 discussion topics on online bulletin board. We analyzed it using language networks, one of the text mining techniques. As a result, 62.5% of students spoke the future of the AI society positively. Second, if there is a self-driving car accident, 39.2% of students thought it is the vehicle owner's responsibility at the current level of autonomous driving. Third, invasion of privacy, abuse of technology, and unbalanced information acquisition were cited as dysfunctions of the development of AI. It was mentioned that ethical education for both AI users and developers is required as a way to minimize malfunctions, and institutional preparations should be carried out in parallel. Fourth, only 19.2% of students showed a positive opinion about a society where face recognition technology is universal. Finally, there was a common opinion that when collecting data including personal information, only the part with the consent should be used. Regarding the use of AI without moral standards, they emphasized the ethical literacy of both users and developers. This study is meaningful in that it provides information necessary to design the contents of artificial intelligence ethics education in liberal arts education.

• Journal of Broadcast Engineering
• /
• v.25 no.3
• /
• pp.362-373
• /
• 2020

Recently, as autonomous vehicles and intelligent CCTV are growing more interest, the efficient object detection is essential technique. The DPM(Deformable Part Models) which is basis of this paper have used a typical object system that represents highly variable objects using mixtures of deformable part for object. Although it shows high detection performance by capturing part shape and configuration of object model, but it is limited to use in real application due to the complicated algorithm. In this paper, instead of image feature pyramid that takes up a large amount of computation in one part of the detector, we propose a method to reduce the computation speed by reconstructing a new image feature pyramid that uses adaptive bilinear interpolation of feature maps obtained on a specific image scale. As a result, the detection performance for object was lowered a little by 2.82%, however, the proposed detection method improved the speed performance by 10% in comparison with original DPM.

• Journal of Convergence for Information Technology
• /
• v.11 no.10
• /
• pp.45-52
• /
• 2021

In this paper, we proposed an Atrous Residual U-Net (AR-UNet) to improve the segmentation accuracy of semantic segmentation method based on U-Net. The U-Net is mainly used in fields such as medical image analysis, autonomous vehicles, and remote sensing images. The conventional U-Net lacks extracted features due to the small number of convolution layers in the encoder part. The extracted features are essential for classifying object categories, and if they are insufficient, it causes a problem of lowering the segmentation accuracy. Therefore, to improve this problem, we proposed the AR-UNet using residual learning and ASPP in the encoder. Residual learning improves feature extraction ability and is effective in preventing feature loss and vanishing gradient problems caused by continuous convolutions. In addition, ASPP enables additional feature extraction without reducing the resolution of the feature map. Experiments verified the effectiveness of the AR-UNet with Cityscapes dataset. The experimental results showed that the AR-UNet showed improved segmentation results compared to the conventional U-Net. In this way, AR-UNet can contribute to the advancement of many applications where accuracy is important.

• Journal of the Korea Convergence Society
• /
• v.12 no.10
• /
• pp.55-61
• /
• 2021

In this paper, we proposed a DeepLabv3+ based encoder-decoder model utilizing an attention mechanism for precise semantic segmentation. The DeepLabv3+ is a semantic segmentation method based on deep learning and is mainly used in applications such as autonomous vehicles, and infrared image analysis. In the conventional DeepLabv3+, there is little use of the encoder's intermediate feature map in the decoder part, resulting in loss in restoration process. Such restoration loss causes a problem of reducing segmentation accuracy. Therefore, the proposed method firstly minimized the restoration loss by additionally using one intermediate feature map. Furthermore, we fused hierarchically from small feature map in order to effectively utilize this. Finally, we applied an attention mechanism to the decoder to maximize the decoder's ability to converge intermediate feature maps. We evaluated the proposed method on the Cityscapes dataset, which is commonly used for street scene image segmentation research. Experiment results showed that our proposed method improved segmentation results compared to the conventional DeepLabv3+. The proposed method can be used in applications that require high accuracy.