• Title/Summary/Keyword: IoU

Search Result 192, Processing Time 0.032 seconds

Road Extraction from Images Using Semantic Segmentation Algorithm (영상 기반 Semantic Segmentation 알고리즘을 이용한 도로 추출)

  • Oh, Haeng Yeol;Jeon, Seung Bae;Kim, Geon;Jeong, Myeong-Hun
    • Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
    • /
    • v.40 no.3
    • /
    • pp.239-247
    • /
    • 2022
  • Cities are becoming more complex due to rapid industrialization and population growth in modern times. In particular, urban areas are rapidly changing due to housing site development, reconstruction, and demolition. Thus accurate road information is necessary for various purposes, such as High Definition Map for autonomous car driving. In the case of the Republic of Korea, accurate spatial information can be generated by making a map through the existing map production process. However, targeting a large area is limited due to time and money. Road, one of the map elements, is a hub and essential means of transportation that provides many different resources for human civilization. Therefore, it is essential to update road information accurately and quickly. This study uses Semantic Segmentation algorithms Such as LinkNet, D-LinkNet, and NL-LinkNet to extract roads from drone images and then apply hyperparameter optimization to models with the highest performance. As a result, the LinkNet model using pre-trained ResNet-34 as the encoder achieved 85.125 mIoU. Subsequent studies should focus on comparing the results of this study with those of studies using state-of-the-art object detection algorithms or semi-supervised learning-based Semantic Segmentation techniques. The results of this study can be applied to improve the speed of the existing map update process.

Waterbody Detection Using UNet-based Sentinel-1 SAR Image: For the Seom-jin River Basin (UNet기반 Sentinel-1 SAR영상을 이용한 수체탐지: 섬진강유역 대상으로)

  • Lee, Doi;Park, Soryeon;Seo, Dongju;Kim, Jinsoo
    • Korean Journal of Remote Sensing
    • /
    • v.38 no.5_3
    • /
    • pp.901-912
    • /
    • 2022
  • The frequency of disasters is increasing due to global climate change, and unusual heavy rains and rainy seasons are occurring in Korea. Periodic monitoring and rapid detection are important because these weather conditions can lead to drought and flooding, causing secondary damage. Although research using optical images is continuously being conducted to determine the waterbody, there is a limitation in that it is difficult to detect due to the influence of clouds in order to detect floods that accompany heavy rain. Therefore, there is a need for research using synthetic aperture radar (SAR) that can be observed regardless of day or night in all weather. In this study, using Sentinel-1 SAR images that can be collected in near-real time as open data, the UNet model among deep learning algorithms that have recently been used in various fields was applied. In previous studies, waterbody detection studies using SAR images and deep learning algorithms are being conducted, but only a small number of studies have been conducted in Korea. In this study, to determine the applicability of deep learning of SAR images, UNet and the existing algorithm thresholding method were compared, and five indices and Sentinel-2 normalized difference water index (NDWI) were evaluated. As a result of evaluating the accuracy with intersect of union (IoU), it was confirmed that UNet has high accuracy with 0.894 for UNet and 0.699 for threshold method. Through this study, the applicability of deep learning-based SAR images was confirmed, and if high-resolution SAR images and deep learning algorithms are applied, it is expected that periodic and accurate waterbody change detection will be possible in Korea.

A Study on Automated Stock Trading based on Volatility Strategy and Fear & Greed Index in U.S. Stock Market (미국주식 매매의 변동성 전략과 Fear & Greed 지수를 기반한 주식 자동매매 연구)

  • Sunghyuck Hong
    • Advanced Industrial SCIence
    • /
    • v.2 no.3
    • /
    • pp.22-28
    • /
    • 2023
  • In this study, we conducted research on the automated trading of U.S. stocks through a volatility strategy using the Fear and Greed index. Volatility in the stock market is a common phenomenon that can lead to fluctuations in stock prices. Investors can capitalize on this volatility by implementing a strategy based on it, involving the buying and selling of stocks based on their expected level of volatility. The goal of this thesis is to investigate the effectiveness of the volatility strategy in generating profits in the stock market.This study employs a quantitative research methodology using secondary data from the stock market. The dataset comprises daily stock prices and daily volatility measures for the S&P 500 index stocks. Over a five-year period spanning from 2016 to 2020, the stocks were listed on the New York Stock Exchange (NYSE). The strategy involves purchasing stocks from the low volatility group and selling stocks from the high volatility group. The results indicate that the volatility strategy yields positive returns, with an average annual return of 9.2%, compared to the benchmark return of 7.5% for the sample period. Furthermore, the findings demonstrate that the strategy outperforms the benchmark return in four out of the five years within the sample period. Particularly noteworthy is the strategy's performance during periods of high market volatility, such as the COVID-19 pandemic in 2020, where it generated a return of 14.6%, as opposed to the benchmark return of 5.5%.

Development of Deep Learning Structure for Defective Pixel Detection of Next-Generation Smart LED Display Board using Imaging Device (영상장치를 이용한 차세대 스마트 LED 전광판의 불량픽셀 검출을 위한 딥러닝 구조 개발)

  • Sun-Gu Lee;Tae-Yoon Lee;Seung-Ho Lee
    • Journal of IKEEE
    • /
    • v.27 no.3
    • /
    • pp.345-349
    • /
    • 2023
  • In this paper, we propose a study on the development of deep learning structure for defective pixel detection of next-generation smart LED display board using imaging device. In this research, a technique utilizing imaging devices and deep learning is introduced to automatically detect defects in outdoor LED billboards. Through this approach, the effective management of LED billboards and the resolution of various errors and issues are aimed. The research process consists of three stages. Firstly, the planarized image data of the billboard is processed through calibration to completely remove the background and undergo necessary preprocessing to generate a training dataset. Secondly, the generated dataset is employed to train an object recognition network. This network is composed of a Backbone and a Head. The Backbone employs CSP-Darknet to extract feature maps, while the Head utilizes extracted feature maps as the basis for object detection. Throughout this process, the network is adjusted to align the Confidence score and Intersection over Union (IoU) error, sustaining continuous learning. In the third stage, the created model is employed to automatically detect defective pixels on actual outdoor LED billboards. The proposed method, applied in this paper, yielded results from accredited measurement experiments that achieved 100% detection of defective pixels on real LED billboards. This confirms the improved efficiency in managing and maintaining LED billboards. Such research findings are anticipated to bring about a revolutionary advancement in the management of LED billboards.

Effective Multi-Modal Feature Fusion for 3D Semantic Segmentation with Multi-View Images (멀티-뷰 영상들을 활용하는 3차원 의미적 분할을 위한 효과적인 멀티-모달 특징 융합)

  • Hye-Lim Bae;Incheol Kim
    • KIPS Transactions on Software and Data Engineering
    • /
    • v.12 no.12
    • /
    • pp.505-518
    • /
    • 2023
  • 3D point cloud semantic segmentation is a computer vision task that involves dividing the point cloud into different objects and regions by predicting the class label of each point. Existing 3D semantic segmentation models have some limitations in performing sufficient fusion of multi-modal features while ensuring both characteristics of 2D visual features extracted from RGB images and 3D geometric features extracted from point cloud. Therefore, in this paper, we propose MMCA-Net, a novel 3D semantic segmentation model using 2D-3D multi-modal features. The proposed model effectively fuses two heterogeneous 2D visual features and 3D geometric features by using an intermediate fusion strategy and a multi-modal cross attention-based fusion operation. Also, the proposed model extracts context-rich 3D geometric features from input point cloud consisting of irregularly distributed points by adopting PTv2 as 3D geometric encoder. In this paper, we conducted both quantitative and qualitative experiments with the benchmark dataset, ScanNetv2 in order to analyze the performance of the proposed model. In terms of the metric mIoU, the proposed model showed a 9.2% performance improvement over the PTv2 model using only 3D geometric features, and a 12.12% performance improvement over the MVPNet model using 2D-3D multi-modal features. As a result, we proved the effectiveness and usefulness of the proposed model.

Competition and Host-strain Interaction of Soybean Rhizobium Strains on Two Soybean Cultivars (콩 근류균계간 경합과 숙주 친화성의 품종간 차이)

  • 박의호;싱글톤폴
    • KOREAN JOURNAL OF CROP SCIENCE
    • /
    • v.41 no.6
    • /
    • pp.718-724
    • /
    • 1996
  • Two soybean cultivars, ‘Lee’ and ‘Peking’, were used to evaluate the competition and interaction of rhizobium strains PRC205 (R. fredii, fast-grower) and USDA110 (B. japonicum, slow-grower). Strains were inoculated separately on the root parts of a split-root growth system. Both root sides were inoculated simultaneously with four combinations of strain treatment to evaluate the competition of strains. And to evaluate the interaction of strains one side of split-root system was inoculated a week prior to the other side. Nodule mass and dry weight of the plants were measured 3 weeks after treatments. PRC205 showed no effective nodulation and no competing ability with USDA110 on Lee cultivar, however, contrary results on Peking cultivar. Top dry weight of Lee inoculated with PRC205 was much lower than that of any other inoculation treatments, however, in Peking that with PRC205 was higher than that with USDA110. There were no differences in root dry weight among the inoculation treatments. USDA110 used as primary inoculant suppressed nodule mass of opposite side, secondary inoculant, severely in both cultivars. PRC205 showed same tendency as USDA110 in Peking, but revealed little suppression effects on USDA110 used as secondary inoculant in Lee. USDA110 used as primary inoculant in Lee and PRC205 in Peking showed much more dry weight of soybean plants than that of other treatments.

  • PDF

Laser Damage Threshold Increase of A/R Coating Films for 200MHz AOM (A/R 코팅 변화에 따른 200MHz AOM의 laser damage threshold 증가)

  • Kim, Yong-Hun;Lee, Hang-Hun;Lee, Jin-Ho;Park, Yeong-Jun;Park, Jeong-Ho
    • Korean Journal of Materials Research
    • /
    • v.7 no.3
    • /
    • pp.213-217
    • /
    • 1997
  • AOhf(Hcousto-r)l)tic niodulator) with :!OOlIiz freclucncl- and Sfi(;(Seconrl harmonic generation) green lasel-Lvith 53% nm wavelength were used for Il\'IIII~Dii.it,ii v~ilco disk recorder) FOI rhe appli~aptin of high densit]. optical recording, a high po\ver I ~ c r is r c ~ ~ l i ~ i l - u l ic I !tic. s\-sti,m a n d optic.,~I io;iting l,t)c>rs of each optical device must have a high laser damage threshoid hie rn;itie ant] retlwtive coatings on a $TeO_{2}$ singlc crystal. which is used as an acoustooptic material, by E-beam evaporation method. Laser damage threshold \vas nicdsureci hy Ar laser with the input power oi 0.55LV 1,aser damage threiholti 01 $ZrO_{2}$ and $SiO_{2}$. filn-is were higher than $AI_{2}O_{3}$ f i l m U'e also investigated a long--tern1 stability of the output po\ver of St{(; green laser

  • PDF

A Study on the Derivation of isoflux Radiation Characteristics of Planar Array Antenna for CubeSAT (큐브위성용 평면배열안테나의 isoflux 방사특성 유도에 관한 연구)

  • Jung, Jinwoo;Pyo, Seongmin
    • Journal of IKEEE
    • /
    • v.24 no.3
    • /
    • pp.917-920
    • /
    • 2020
  • In this paper, we studied the derivation method of isoflux radiation for a planar array antenna. The presented array antenna was designed for considering of 1U-sized CubeSAT with Ku-band communications. For the presented array antenna, 8×8 radiating elements were arranged, and the distance between radiating elements was set of half-wavelength. The excited current weighting for each radiating elements was calculated by the signal processing technique used in the design of the low-pass filter. As a result of analysis of the method proposed in this paper, it was confirmed that a seamless isoflux pattern can be derived.

Deconvolution Pixel Layer Based Semantic Segmentation for Street View Images (디컨볼루션 픽셀층 기반의 도로 이미지의 의미론적 분할)

  • Wahid, Abdul;Lee, Hyo Jong
    • Proceedings of the Korea Information Processing Society Conference
    • /
    • 2019.05a
    • /
    • pp.515-518
    • /
    • 2019
  • Semantic segmentation has remained as a challenging problem in the field of computer vision. Given the immense power of Convolution Neural Network (CNN) models, many complex problems have been solved in computer vision. Semantic segmentation is the challenge of classifying several pixels of an image into one category. With the help of convolution neural networks, we have witnessed prolific results over the time. We propose a convolutional neural network model which uses Fully CNN with deconvolutional pixel layers. The goal is to create a hierarchy of features while the fully convolutional model does the primary learning and later deconvolutional model visually segments the target image. The proposed approach creates a direct link among the several adjacent pixels in the resulting feature maps. It also preserves the spatial features such as corners and edges in images and hence adding more accuracy to the resulting outputs. We test our algorithm on Karlsruhe Institute of Technology and Toyota Technologies Institute (KITTI) street view data set. Our method achieves an mIoU accuracy of 92.04 %.

A Study of the Metal Recovery from the Aluminium Scrap (Al 스크랩으로부터 금속회수에 관한 연구)

  • 김준수;임병모;윤의박
    • Resources Recycling
    • /
    • v.4 no.1
    • /
    • pp.25-30
    • /
    • 1995
  • In the preparatIon of reclaimed aluminium lllgot from alumimum scrap, the aluminium recovery was studied a as a function of the preliminary treatment of samples, addition of flux and melting atmosphere. AI dross is produced by an oxidation reaction at the surface of liquid metal. The recovery of AI metal increases u up to maximum 95% by adding salt up to 7%, The recovery of AI metal in the compacted chip bale without oil removal mcrease about 14% compared io non-compacted chip. In the case of the AI seed melting process, the recovery of Al metal of the crushed and compacted chip hale is 97%, In meltmg of alumimum scrap under the atmosphere of carbon and nitrogen gas, the recovery of AI metal increase, but it is decreased when the mixture of salt and carbon powder is added excessively.

  • PDF