• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.17 no.3
• /
• pp.98-110
• /
• 2018

In this paper, we propose both Dense grid model and Anchor model to improve the recognition rate of dynamic objects. Two experiments are conducted to study the performance of two proposed CNNs models (Dense grid model and Anchor model), which are to detect dynamic objects. In the first experiment, YOLO-v2 network is adjusted, and then fine-tuned on KITTI datasets. The Dense grid model and Anchor model are then compared with YOLO-v2. Regarding to the evaluation, the two models outperform YOLO-v2 from 6.26% to 10.99% on car detection at different difficulty levels. In the second experiment, this paper conducted further training of the models on a new dataset. The two models outperform YOLO-v2 up to 22.40% on car detection at different difficulty levels.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2023.07a
• /
• pp.75-78
• /
• 2023

본 논문은 객체탐지 모델 중 주류를 이루고 있는 YOLO의 v1부터 v8까지의 특성을 비교 분석하여 각각의 버전에 최적화할 수 있는 모델에 대한 연구이다. 연구 결과 v1, v2는 정확성이 최우선인 모델에 적합하다. 반면, v3, v4는 속도가 우선인 모델에 적합하다. 또한 v5, v6는 정확도와 속도 사이의 균형이 필요한 모델에 적합하다는 결론을 얻었다. v7, v8은 메모리 및 컴퓨팅 성능에 제약이 있는 경우 주로 적용이 가능하며, 적은 연산과 저 메모리 사용으로 객체를 탐지하여 포즈추정이나 객체 추적 등을 적용할 모델에 적합하다는 결과를 확인하였다.

• International Journal of Advanced Culture Technology
• /
• v.10 no.1
• /
• pp.230-235
• /
• 2022

This study is to select an optimal object detection algorithm for designing a self-checkout counter to improve the inconvenience of payment systems for products without existing barcodes. To this end, a performance comparison analysis of YOLO v2, Tiny YOLO v2, and the latest YOLO v5 among deep learning-based object detection algorithms was performed to derive results. In this paper, performance comparison was conducted by forming learning data as an example of 'donut' in a bakery store, and the performance result of YOLO v5 was the highest at 96.9% of mAP. Therefore, YOLO v5 was selected as the artificial intelligence object detection algorithm to be applied in this paper. As a result of performance analysis, when the optimal threshold was set for each donut, the precision and reproduction rate of all donuts exceeded 0.85, and the majority of donuts showed excellent recognition performance of 0.90 or more. We expect that the results of this paper will be helpful as the fundamental data for the development of an automatic payment system using AI self-service technology that is highly usable in the non-face-to-face era.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.25 no.1
• /
• pp.36-43
• /
• 2021

In this research, the accuracy of YOLO v3 algorithm in object detection during AGV (Automated Guided Vehicle) operation was investigated. First of all, AGV with 2D LiDAR and stereo camera was prepared. AGV was driven along the route scanned with SLAM (Simultaneous Localization and Mapping) using 2D LiDAR while front objects were detected through stereo camera. In order to evaluate the accuracy of YOLO v3 algorithm, recall, AP (Average Precision), and mAP (mean Average Precision) of the algorithm were measured with a degree of machine learning. Experimental results show that mAP, precision, and recall are improved by 10%, 6.8%, and 16.4%, respectively, when YOLO v3 is fitted with 4000 training dataset and 500 testing dataset which were collected through online search and is trained additionally with 1200 dataset collected from the stereo camera on AGV.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.18 no.6
• /
• pp.155-163
• /
• 2019

This paper proposes a new methodology to recognize cracks on asphalt road surfaces using the image data obtained with drones. The target section was Yuseong-daero, the main highway of Daejeon. Furthermore, two object detection algorithms, such as Tiny-YOLO-V2 and Faster-RCNN, were used to recognize cracks on road surfaces, classify the crack types, and compare the experimental results. As a result, mean average precision of Faster-RCNN and Tiny-YOLO-V2 was 71% and 33%, respectively. The Faster-RCNN algorithm, 2Stage Detection, showed better performance in identifying and separating road surface cracks than the Yolo algorithm, 1Stage Detection. In the future, it will be possible to prepare a plan for building an infrastructure asset-management system using drones and AI crack detection systems. An efficient and economical road-maintenance decision-support system will be established and an operating environment will be produced.

• Journal of Advanced Navigation Technology
• /
• v.27 no.6
• /
• pp.776-781
• /
• 2023

K-UAM will be commercialized through maturity after 2035. Since the Urban Air Mobility (UAM) corridor will be used vertically separating the existing helicopter corridor, the corridor usage is expected to increase. Therefore, a system for monitoring corridors is also needed. In recent years, object detection algorithms have developed significantly. Object detection algorithms are largely divided into one-stage model and two-stage model. In real-time detection, the two-stage model is not suitable for being too slow. One-stage models also had problems with accuracy, but they have improved performance through version upgrades. Among them, YOLO-V5 improved small image object detection performance through Mosaic. Therefore, YOLO-V5 is the most suitable algorithm for systems that require real-time monitoring of wide corridors. Therefore, this paper trains YOLO-V5 and analyzes whether it is ultimately suitable for corridor monitoring.K-uam will be commercialized through maturity after 2035.

• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.12 no.5
• /
• pp.467-474
• /
• 2019

In this paper, we propose a novel method to improve FPS while maintaining the accuracy of YOLO v2 model in NVIDIA Jetson TX1. In general, in order to reduce the amount of computation, a conversion to an integer operation or reducing the depth of a network have been used. However, the accuracy of recognition can be deteriorated. So, we use methods to reduce computation and memory consumption through adjustment of the filter size and integrated computation of the network The first method is to replace the $3{\times}3$ filter with a $1{\times}1$ filter, which reduces the number of parameters to one-ninth. The second method is to reduce the amount of computation through CBR (Convolution-Add Bias-Relu) among the inference acceleration functions of TensorRT, and the last method is to reduce memory consumption by integrating repeated layers using TensorRT. For the simulation results, although the accuracy is decreased by 1% compared to the existing YOLO v2 model, the FPS has been improved from the existing 3.9 FPS to 11 FPS.

• Smart Media Journal
• /
• v.11 no.4
• /
• pp.46-55
• /
• 2022

It is very important to first detect and remove defective fruits with scratches or bruised areas in the automatic fruit quality screening system. This paper proposes a method of detecting defective areas in fruits using the latest artificial intelligence technology, the Yolo V4 deep learning model in order to overcome the limitations of the method of detecting fruit's defective areas using the existing image processing techniques. In this study, a total of 2,400 defective fruits, including 1,000 defective apples and 1,400 defective fruits with scratch or decayed areas, were learned using the Yolo V4 deep learning model and experiments were conducted to detect defective areas. As a result of the performance test, the precision of apples is 0.80, recall is 0.76, IoU is 69.92% and mAP is 65.27%. The precision of pears is 0.86, recall is 0.81, IoU is 70.54% and mAP is 68.75%. The method proposed in this study can dramatically improve the performance of the existing automatic fruit quality screening system by accurately selecting fruits with defective areas in real time rather than using the existing image processing techniques.

• Journal of Broadcast Engineering
• /
• v.24 no.3
• /
• pp.495-505
• /
• 2019

Typical algorithms of deep learning include CNN(Convolutional Neural Networks), which are mainly used for image recognition, and RNN(Recurrent Neural Networks), which are used mainly for speech recognition and natural language processing. Among them, CNN is able to learn from filters that generate feature maps with algorithms that automatically learn features from data, making it mainstream with excellent performance in image recognition. Since then, various algorithms such as R-CNN and others have appeared in object detection to improve performance of CNN, and algorithms such as YOLO(You Only Look Once) and SSD(Single Shot Multi-box Detector) have been proposed recently. However, since these deep learning-based detection algorithms determine the success of the detection in the still images, stable object tracking and detection in the video requires separate tracking capabilities. Therefore, this paper proposes a method of combining Kalman filters into deep learning-based detection networks for improved object tracking and detection performance in the video. The detection network used YOLO v2, which is capable of real-time processing, and the proposed method resulted in 7.7% IoU performance improvement over the existing YOLO v2 network and 20 fps processing speed in FHD images.

• Journal of Korean Society of Rural Planning
• /
• v.28 no.2
• /
• pp.87-96
• /
• 2022

As part of strengthening energy security and responding to climate change, the government has promoted various renewable energy measures to increase the development of renewable energy facilities. As a result, small-scale solar installations in rural areas have increased rapidly. The number of complaints from local residents is increasing. Therefore, in this study, deep learning technology is applied to high-resolution aerial images on the internet to detect solar power plants installed in rural areas to determine whether or not solar power plants are installed. Specifically, I examined the solar facility detector generated by training the YOLO(You Only Look Once) v2 object detector and looked at its usability. As a result, about 800 pieces of training data showed a high object detection rate of 93%. By constructing such an object detection model, it is expected that it can be utilized for land use monitoring in rural areas, and it can be utilized as a spatial data construction plan for rural areas using technology for detecting small-scale agricultural facilities.