본 연구는 오픈소스 라이브러리인 OpenCV를 활용해 다양 한 시설과채류의 표현형 분석에 적용 가능한 컴퓨터 비전 기 술을 탐구하였다. 토마토에 대해서는 이미지의 색상을 분석 하여 숙성도를 판정하며, support vector machine(SVM) and histogram of oriented gradients 기법을 통해 숙성된 토마토 를 효과적으로 검출하였다. 파프리카의 경우, 색상 분포를 시 각화한 후, 가우스 혼합 모델로 클러스터링을 실행하여 수확 파프리카의 색상 특성을 분석하였다. 네트 멜론의 품질 평가 에서는 LAB 색상 공간, 이진화 이미지 및 깊이 매핑을 활용하 여 멜론의 네트 패턴을 정량화하였다. 추가로, 오이 온실에서 화방 검출을 위해 깊이 정보와 색상 정보를 조합하여 다양한 크기와 거리의 화방을 성공적으로 검출하였다. 이 연구의 결 과로, 해당 컴퓨터 비전 기술들이 시설과채류의 생장 모니터 링, 숙성 및 품질 평가 등에서의 유효성을 확인하였다. 농산업 에서 컴퓨터 비전의 효과적 적용을 위해, 후속 연구자나 개발 자들이 재배 생리와 연관된 지표를 기반으로 이 기술들을 보 완할 경우, 실제 농업 현장 및 연구에서 널리 활용될 가능성이 크다.

With the recent surge in YouTube usage, there has been a proliferation of user-generated videos where individuals evaluate cosmetics. Consequently, many companies are increasingly utilizing evaluation videos for their product marketing and market research. However, a notable drawback is the manual classification of these product review videos incurring significant costs and time. Therefore, this paper proposes a deep learning-based cosmetics search algorithm to automate this task. The algorithm consists of two networks: One for detecting candidates in images using shape features such as circles, rectangles, etc and Another for filtering and categorizing these candidates. The reason for choosing a Two-Stage architecture over One-Stage is that, in videos containing background scenes, it is more robust to first detect cosmetic candidates before classifying them as specific objects. Although Two-Stage structures are generally known to outperform One-Stage structures in terms of model architecture, this study opts for Two-Stage to address issues related to the acquisition of training and validation data that arise when using One-Stage. Acquiring data for the algorithm that detects cosmetic candidates based on shape and the algorithm that classifies candidates into specific objects is cost-effective, ensuring the overall robustness of the algorithm.

Melon fruits exhibit a wide range of morphological variations in fruit shape, sugar content, net quality, diameter and weight, which are largely dependent on the variety. These characteristics significantly affect marketability. For netted varieties, the uniformity and pattern of the net serve as key factors in determining the external quality of the melon and act as indicators of its internal quality. In this study, we evaluated the effect of fruit morphology and growth on netting by analyzing the changes in melon fruit quality under LED light treatment and monitoring fruit growth. Computer vision analysis was used for quantitative evaluation of fruit net quality, and a three-variable logistic model was applied to simulate fruit growth. The results showed that melons grown under LED conditions exhibited more uniform fruit shape and improvements in both net quality and sugar content compared to the control group. The results of the logistic model showed minimal error values and consistent curve slopes across treatments, confirming its ability to accurately predict fruit growth patterns under varying light conditions. This study provides an understanding of the effects of fruit shape and growth on net quality.

In the realm of dental prosthesis fabrication, obtaining accurate impressions has historically been a challenging and inefficient process, often hindered by hygiene concerns and patient discomfort. Addressing these limitations, Company D recently introduced a cutting-edge solution by harnessing the potential of intraoral scan images to create 3D dental models. However, the complexity of these scan images, encompassing not only teeth and gums but also the palate, tongue, and other structures, posed a new set of challenges. In response, we propose a sophisticated real-time image segmentation algorithm that selectively extracts pertinent data, specifically focusing on teeth and gums, from oral scan images obtained through Company D's oral scanner for 3D model generation. A key challenge we tackled was the detection of the intricate molar regions, common in dental imaging, which we effectively addressed through intelligent data augmentation for enhanced training. By placing significant emphasis on both accuracy and speed, critical factors for real-time intraoral scanning, our proposed algorithm demonstrated exceptional performance, boasting an impressive accuracy rate of 0.91 and an unrivaled FPS of 92.4. Compared to existing algorithms, our solution exhibited superior outcomes when integrated into Company D's oral scanner. This algorithm is scheduled for deployment and commercialization within Company D's intraoral scanner.

본 논문은 AI와 시각, 컴퓨터비전과 이미지의 관계를 비평적 관점에서 논하고 있다. 초대형 IT기업들은 빠짐없이 AI의 기계학습을 위한 컴퓨터비전의 개발에 앞장서면서 이미지의 입력을 통해 세상의 정보를 데이터베이스의 형태로 집적하고 있다. 컴퓨터비전의 분야는 ‘비전’이라는 단어가 가리키듯 인간의 시각과 컴퓨터의 시각데이터집적과정을 비유하지만, 실제로는 기계적이고 알고리즘화된 인간의 시각과는 전혀 별개의 성질을 갖는다. 기계에 정보를 제공하는 기계학 습을 위한 이미지의 모음인 이미지 데이터셋은 AI의 성능개발에 핵심적인데, 온라인상의 이미지가 무작위로 사용되거나 학습과정에서 사회적 차별이나 편견이 그대로 반영될 가능성이 높은 것 으로 경계가 필요하다. AI와 예술의 접목에서는 주로 생성적 적대 신경망을 사용하여 기존의 미술품을 학습한 후 이와 유사하지만 다른 이미지를 만들어내는 방식이 다용되고 있다. 컴퓨터비전 이 시각과 차이가 있듯, AI 미술이 기존 예술의 일부로 흡수되기 위해서는 새로운 기준들이 필요할 것이다.

PURPOSES : This study aims to develop and evaluate computer vision-based algorithms that classify the road roughness index (IRI) of road specimens with known IRIs. The presented study develops and compares classifier-based and deep learning-based models that can effectively determine pavement roughness grades. METHODS : A set road specimen was developed for various IRIs by generating road profiles with matching standard deviations. In addition, five distinct features from road images, including mean, peak-to-peak, standard variation, and mean absolute deviation, were extracted to develop a classifier-based model. From parametric studies, a support vector machine (SVM) was selected. To further demonstrate that the model is more applicable to real-world problems, with a non-integer road grade, a deep-learning model was developed. The algorithm was proposed by modifying the MNIST database, and the model input parameters were determined to achieve higher precision. RESULTS : The results of the proposed algorithms indicated the potential of using computer vision-based models for classifying road surface roughness. When SVM was adopted, near 100% precision was achieved for the training data, and 98% for the test data. Although the model indicated accurate results, the model was classified based on integer IRIs, which is less practical. Alternatively, a deep-learning model, which can be applied to a non-integer road grade, indicated an accuracy of over 85%. CONCLUSIONS : In this study, both the classifier-based, and deep-learning-based models indicated high precision for estimating road surface roughness grades. However, because the proposed algorithm has only been verified against the road model with fixed integers, optimization and verification of the proposed algorithm need to be performed for a real road condition.

목적 : 컴퓨터를 많이 사용하는 대학생을 대상으로 온라인 학습 환경과 CVS 자각증상을 조사하여 CVS 증상 예방조치를 알아보고자 하였다. 방법 : 2021년 3월부터 4월까지 온라인 설문조사에 동의한 대학생 140명을 대상으로 하였다. 대상자들의 평균 연령은 19.87±1.48세로 남학생과 여학생은 각각 70명(50.0%) 이었다. 연구 도구는 대상자의 일반적인 특성(4), 온라인 학습 환경 특성(12), CVS 자각증상(16) 총 32문항의 설문지를 이용하였다. 결과 : 대상자의 CVS 증상은 ‘건조함’(69.3%)이 가장 높은 빈도로 나타났으며, ‘시력이 나빠지는 느낌’(62.9%), ‘두통’(60.7%), ‘눈의 통증’(53.6%), ‘눈꺼풀 무거움’(50.7%)순으로 많이 나타났다. 컴퓨터 사용시간이 8시간 이상 인 그룹의 CVS 증상 발생 빈도가 높았으며, 온라인 학습동안 규칙적으로 휴식하지 않는 그룹이 휴식을 하는 그룹 보다 모든 CVS 증상에서 발생 빈도가 높았고 통계적으로 유의하였다. 결론 : 온라인 학습 환경과 CVS 자각증상을 확인하였다. 컴퓨터 사용 시간이 길수록 CVS 증상의 발생빈도는 높았으며, 온라인 학습동안 규칙적으로 휴식하는 것이 CVS 증상의 빈도를 감소시키는 것을 확인하였고, 규칙적인 눈 운동 필요 인식이 높은 결과를 확인하였다. 따라서 온라인 학습동안 올바른 눈 운동과 휴식 방법의 홍보와 교육이 필요하다고 생각한다.

컴퓨터 비젼을 이용한 항행선박의 항적을 계산하고 교통량을 측정하는 방법은 해양사고의 예방관점에서 사고발생 가능성 여부를 예측해 볼 수 있는 유용한 방법이다. 본 연구에서는 컴퓨터 비젼을 이용하여 영상축소, 미분연산자, 최대 최소값 등을 이용하여 선박을 인식한 후 실세계 상에서의 좌표 값을 계산하여 실시간 항적을 전자 해도에 표시함으로서 해상 구조물과의 충돌여부를 직접 육안으로 확인 할 수 있는 알고리즘을 개발하였다. 본 연구에서 개발된 알고리즘은 영역 정보를 기반으로 개발되었기 때문에 점 정보에 의존하고 있는 기존 레이더 시스템의 단점을 보완하는 장점을 지니고 있다.

철도교의 장기변위 정보는 시공 및 유지관리에 있어 매우 유용하지만, 실구조물의 장기간에 걸쳐 발생하는 변위를 정확하게 계측하기 위해서는 많은 실질적인 문제를 해결해야 한다. 본 연구에서는 철도교량의 효과적인 장기변위 계측을 위해, 컴퓨터 비전 기반의 비접촉식 기법을 제안한다. 컴퓨터 비전 기반 기법은 비용적인 측면에서 우수하며 사용이 간편하여 최근 교량변위 계측을 위해 기술개발이 활발하게 되고 있으나, 카메라의 미소변위에 의해 큰 오차가 발생하므로, 장기변위의 계측에는 적합하지 않다. 본 연구에서는 두 개의 카메라를 이용하여 카메라 변위에 따른 오차를 보정하는 방식으로 장기변위 계측을 가능하게 하였다. 개발된 기법을 시공 중인 철도교량에 적용하여 성능을 검증하였다.

The camera movement can cause considerable error in the computer vision-based displacement measurement because the camera is generally placed far from the region of interest. The camera movement, which is difficult to avoid particularly in the long-term measurement, hinders real-world applications of the computer vision-based displacement method. This research proposes a practical means of long-term displacement measurement by using a novel camera system. In addition to the conventional camera-based measurement, an auxiliary camera is used to compensate the camera motion-induced error. Experimental validation is conducted in the laboratory environment.

recently, information about buried objects has been needed for redevelopment and reorganization of the complicated urban environment. Accidents caused by pipeline damages, such as gas lines, communication lines and underground electric power lines, are results of loss of people and property. Therefore, information on underground obscured material is essential for safety and construction progress. GPR (Ground Penetrating Radar) investigation has advantages of high resolution, ease of utilization and strong electromagnetic noise when using high frequency. However, the GPR detection data image is not visible and has a problem that it is interpreted differently according to the skill of the inspector. Therefore, this study was conducted to verify the visualization of detection data using computer vision based on GPR detection data. Canny edge and Harris corner detection were applied to the GPR image data to detect the hyperbolic shape. By using this to increase the visibility, it will contribute to the reliable result in the buried detection.

Computer vision-based displacement measurement is regarded as an excellent alternative to conventional displacement measurement devices due to its convenient installation process with high accuracy. Based on the strong potential, this study proposes a computer vision approach for displacement measurement with enhanced field applicability. Main features of the proposed method are (1) robustness against adverse light conditions and (2) convenient camera installation that allows the camera to be arbitrarily placed. An adaptive image processing procedure is developed to overcome false identification of target markers that can be induced by strong lights. Location information of the identified markers are used to obtain displacement using the homography transformation, which allows the camera to be installed in any place as long as the target markers are in the line of sight. With these features, the computer vision approach is experimentally proven to be practical in field testing for measuring structural displacement.

SHM에서 변위는 구조물의 동적 특성을 파악하는 핵심정보다. 이를 구조물과 직접 접촉하여 자기장 변화를 전기적신호(LVDT), 직접적으로 측정하는 방안으로 LVDT, GPS, LDV Displacement is one of the most fundamental responses, containing useful information regarding dynamic behavior of a structure. Traditional displacement measurement devices such as LVDT have disadvantages in its high-cost and few options on installation place. For the sake of economic reason, vision-based displacement measurement systems using low-cost cameras have been developed, yet these approaches still have difficulties in finding appropriate camera positions; camera should be placed perpendicular to targets. This study presents a new vision-based displacement measurement system using the planar homography method that gives accurate displacement, allowing the camera to have an arbitrary angle toward target. The vision-based system is experimentally verified using a low-cost camera and a target with four circles.

Camera arrangement for depth and image correspondence is very important to the computer vision. Two conventional comera arrangements for stereo computer vision are lateral model and axial motion model. In this paper, using the axial motion stereo camera model, the algorithm for camera focal length measurement and the surface smoothness with the radiance-irradiance is proposed fro 3-dimensional image correspondence on stereo computer vision. By adapting the above algorithm, camera focal length can be measured precisely and the resolution of 3-dimensional image correspondence has been improved comparing to that of the axial motion model without the radiance-irradiance relation.