Earticle

In this study, we propose a deep learning-based forest road recognition technology and integrate it into an autonomous steering and speed control system for a forestry forwarder. We trained a YOLOv8 segmentation model on 640×480 pixel images collected from straight, curved, and occluded sections of forestry roads, achieving a precision of 0.997, a recall of 0.994, an F1 score of 0.995, and an average precision of 0.994. We developed a centerline extraction algorithm that calculates the lateral offset between the vehicle center and the detected road centerline using the segmented mask. This offset is then input into an electronic steering lever control system, which replaces the existing mechanical steering lever to enable rapid response. In field experiments conducted on actual forest roads, the maximum average path tracking error was 0.534 m over three trials, confirming that the vision-based approach can reliably guide the forwarder under various lighting conditions, shadows, and roadside structures. This method is expected to overcome the limitations of GNSS/INS-based autonomous driving systems, as it operates solely with cameras and single-board hardware. The research results suggest that deep learning-based image processing provides a solid foundation for fully autonomous forestry operations and has the potential to improve the productivity and safety of timber transport operations.

We report processing technique in the MO image measurement system. Calibration procedure is not only considered to perpendicular field but also in-plane field. Current density and field profiles are obtained by Biot-savart law and inversion method. We show example of (Gd,Y)1Ba2Cu3O7-δ-BaZrO3 film that have tilted nano rod pinning centers about 13° from the c-axis.

본 논문에서는 위성영상처리 소프트웨어 및 시스템의 재사용성을 높이고, 개발기간 및 유지 관리 비용을 줄일 수 있는 알고리즘컴포넌트화를 통한 위성영상처리 프레임워크 및 시스템에 대한 구조를 제안한다. 기존 위성영상처리 소프트웨어 및 시스템은 특정데이터 및 기능에 국한되어 제한된 구조를 갖고 있다. 또한, 각각의 시스템들은 동일 및 유사한 영상처리 알고리즘이 사용되지만해당 알고리즘을 중복적으로 개발하는 문제점이 있다. 이와 같은 문제점을 해결하기 위해 위성영상처리 소프트웨어 프레임워크의요구사항을 분석하였다. 요구사항을 반영한 프레임워크 및 시스템 구조를 설계하였으며, 위성영상처리 프레임워크의 운영흐름도 함께 도출하였다.

This paper suggest meteorological satellite processing software’s structure that reduces time and efforts of modification/upgrade. This structure’s key feature is “algorithm component” that works within framework and eventually to a completeMeteorological satellite processing system. Most of existing Meteorological satellite system is designed around specificfunction and data sets which limits range of modification and upgrade. In addition, re-use of current algorithms becomedifficult although re-use of similar algorithm is the case in many occasions. This inefficiency can be resolved by designinga new framework as a result of detail analysis of collected requirements. A new framework and system architecture has beendesigned. In addition, operational flow of Satellite image processing framework has been described.

21세기에 들어와 국내에 초고속 인터넷의 보급으로 현재 초고속 인터넷 보급률이 전 세계에서 최 상위권에 진입해 있는 상황에서, 과거 전화 접속 네트워크로 인터넷을 사용하던 때 보다 웹 사이트 및 웹 서비스가 활발해지고 있으며, 그에 따른 다양한 서비스들이 지원되고 있다. 과거 PC에서 구동되던 응용 프로그램 군들의 설치가 따로 필요 없이 웹 브라우저를 이용하여 다양한 응용프로그램들이 수행되고 있다. 본 연구는 이러한 현상을 반영하여 다양한 웹 브라우저 환경에서 영상처리를 쉽게 접근하고 사용할 수 있는 영상처리 애플리케이션을 개발하는 것에 중점을 두었다.

In 21st century, our environment of internet is the highest ranking group in the world with the spread of the information superhighway network. Compare to the previous internet environment with the telephone line, the current Web sites and Web services are very active, and diverse services are supported. A great variety of application programs using Web browsers are performed without installation of application programs on personal computer. With the reflection of this state, in this paper, we suggest a new image processing application system that is executable in diverse Web browser environments.

In recent year, many imaging systems have been developed, and it became increasingly important to exchange image data through the computer network. Therefore, it is required to reproduce color image independently on each imaging device. However, even if the image are same, perceived color is not always same under different viewing conditions. On the other hand, even if the image are different, we want to perceive same color under different viewing conditions. Therefore we must know the spectral reflectance information of object. We measured many reflectance human skin can be estimate using only three principal component. For Munsell color patches, five principle components were necessary to estimate the reflectance spectra. For that purpose, we have developed color image acquisition system that is composed of five band filters and CCD camera. Improved spectral reflectance of object is predicted by five band images taken by color image acquisition system and then we take account of camera's noise and component of object image for predicting accurate spectral reflectance of object. In the results, we confirmed that color difference and MSE(Mean Square Error) between measured and predicted spectral reflectance of object decreased into 0.0071 and 7.72 respectively.

Tracking a golf ball to visualize its trajectory had previously been done with complex and expensive systems. Those systems are not in the reach of every people. Here we suggest a more straightforward approach for the same purpose but utilizing only golf shots videos. The proposed system uses image processing and a deep neural network model, YOLO (You Only Live Once) V3. To achieve the goal of tracking the golf ball, the YOLO V3 network and Hough transform localize the ball’s initial position and the frame differencing technique tracks the golf ball. Upon implementation of the method to multiple videos, an acceptable result was obtained.

본 연구에서는 고속도로 버스전용차로 단속시스템을 개발하여 현장 성능평가를 수행하였다. 영동고속도로 마성터널 입구 버스전용차로에서 조사한 결과, 버스전용차로를 위반하는 차량의 비율이 99%에 달하는 것으로 조사되었다. 하지만 현재의 경찰관에 의한 인력식 단속은 단속율 도 낮고 불필요한 안전문제 및 지체를 발생시킨다. 고속도로 버스전용차로는 6인 이상 탑승한 9인승 이상 승합차도 통행이 가능하기 때문에 승합차량의 승차인원을 검지하는 기술개발이 필 요하다. 조도에 관계없는 승차인원 검지를 위해 적외선 카메라를 사용하였고 짧은 차두시간을 감안한 신속한 영상처리 기법으로 YOLOv5 딥러닝 알고리즘을 사용하였다. 개발시스템 성능 검증을 위해 테스트베드 및 실 현장 평가를 실시한 결과, 테스트베드와 실 현장에서 각각 7%, 8% 오차를 나타내 만족할 만한 성능을 보였다. 본 연구 결과물을 현장에 적용할 경우 고속도로 버스전용차로 운영 효율화 및 단속에 따른 불필요한 지체를 감소시킬 수 있을 것으로 기대된다.

An automatic freeway bus-only lane enforcement system was developed and assessed in a real-world environment. Observation of a bus-only lane on the Youngdong freeway, South Korea, revealed that approximately 99% of the vehicles violated the high-occupancy vehicle (HOV) lane regulation. However, the current enforcement by the police not only exhibits a low enforcement rate, but also induces unnecessary safety and delay concerns. Since vehicles with six passengers or higher are permitted to enter freeway bus-only lanes, identifying the number of passengers in a vehicle is a core technology required for a freeway bus-only lane enforcement system. To that end, infrared cameras and the You Only Look Once (YOLOv5) deep learning algorithm were utilized. For assessment of the performance of the developed system, two environments, including a controlled test-bed and a real-world freeway, were used. As a result, the performances under the test-bed and the real-world environments exhibited 7% and 8% errors, respectively, indicating satisfactory outcomes. The developed system would contribute to an efficient freeway bus-only lane operations as well as eliminate safety and delay concerns caused by the current manual enforcement procedures.

Modular system refers to the method of pre-configuring the building's base frame, plumbing, insulation and finishing materials in the factory and assembling them on site. In the field assembly, mainly bolted joints are used, and in the event of bolting problems, not only fade the advantages of the modular system, but also adversely affect the joint strength between the modular units. Therefore, in this paper, a system that can detect the manufacturing error of bolt arrangement and analyze the measurement error before the bolt fastening of the modular unit was conducted. An algorithm capable of analyzing image to measure bolt arrangement by receiving images in real time through a camera module that can photograph each modular column and a single board computer was developed. Bolt size, bolt distance, edge distance, modular frame size and perpendicularity were measured. The bolt distance was 0.46mm, 0.22%, and the bolt size was 0.15mm, which was almost accurate with 0.7% error.

본 논문에서는 많은 연산량이 요구되는 영상 처리 알고리즘을 위한 단일 명령어, 다중 명령어 처리 기반 멀티코어프로토타입 시스템을 제안한다. 제안하는 멀티코어 시스템은 16개의 저비용 코어를 가지고 있으며 영상처리 알고리즘에 내재한 데이터 병렬성을 이용하여 성능을 가속화시킨다. 또한 본 논문에서는 제안하는 멀티코어 시스템의 성능을 검증하기 위해 상용 그래픽 프로세싱 유닛 (GPU)과의 실행시간 및 소비 에너지를 비교한다. 동일한1,620MHz 동작 주파수, 16개 코어, 256x256 크기 영상을 이용하여 4단계 화재 감지 알고리즘을 구현한 결과, 제안한 멀티코어 시스템은 상용 GPU (NVIDIA GTX 560)보다 약 1.7배 낮은 실행 시간과 약 88배 낮은 에너지소비를 보였다. 또한 본 논문에서는 제안한 멀티코어 하드웨어를 Virtex-4 FPGA 프로토타입 시스템에 구현하고 기능을 검증하였다.

This paper proposes a single instruction, multiple data (SIMD) based multi-core prototype systemto support high computational image processing algorithms. The proposed multi-core systemincluding 16 low-cost cores accelerates image processing algorithms by exploiting data parallelisminherent in the algorithms. In addition, this paper compares execution time and energyconsumption of the proposed multi-core system and a commercial graphics processing unit to showthe improvement of the proposed multi-core system. Experimental results using the same1,620MHz clock frequency, 16 cores, and a 256x256 image show that a 4-stage fire detectionalgorithm using the multi-core system outperforms that using commercial GPU (NVIDIA GTX 560)in terms of execution time (about 1.7x lower) and energy consumption (about 88x lower). Inaddition, the proposed multi-core hardware was implemented and verified on an FGPA prototypesystem in this paper.

Painting pretreatment is an important task in determining the life of painting as it removes rust or foreign substances from the painting surface and gives adhesion between the painting surface and the painting surface. Since painting pretreatment is an important task, IMO strictly requires that the painting pretreatment surface be maintained at a Sa 2.5 grade and the surface roughness is 30μm~75μm. Painting pre-processing is an important task that determines the lifespan of a painting, but it is done through visual inspection by the inspector, and the quality varies depending on the inspector. In this study, in order to develop a quality measurement system for the painting pretreatment surface, Matlab2023b was used to determine the range of appropriate quality brightness by comparing the brightness of the painting pretreatment surface and surface roughness.

Field programmable gate array (FPGA) has the characteristics of high speed, low power consumption, high integration, flexibility and, small size, etc. In this paper, I design a real time infrared image processing and display system based on FPGA for the requirement of real time infrared image processing, which will realize the transmission, transformation, and storage of the image information, and then complete the infrared image edge detection based on Sobel algorithm by using this system as a platform. At last, the advantages of infrared image processing with this system compared with other methods will be verified in this paper. The experimental results show that the system spends 11.44ms on processing the colorful image whose resolution is 640480.The system has realized the real time, high speed, stable and reliable acquisition, processing and display of infrared image and we can realize the infrared target feature extraction, recognition and tracking when we combine the system with other algorithms.

This study constructed a real-time image processing and object recognition system using MobileNet and evaluated its performance on custom robots. The lightweight architecture of MobileNet enabled high performance in resource-constrained environments, demonstrating its capability for real-time object recognition. The findings validate that a MobileNet-based system is well-suited for optimizing the performance of custom robots, especially in low-power settings. This research highlights the potential of efficient real-time image processing and object recognition implementation to enhance the functionality of robotic systems.

Automatic target-scoring system has great practical value in sports shooting competitions. From the needs of sports shooting competition point of view, this paper used image processing technology and studied a set of sport shooting competition ring number automatic identification system. In this system, the author firstly used the image registration algorithm based on SIFT and achieved accurate registration of two target images before and after shooting. And then OTSU thresholding algorithm was used to eliminate most continuous noise. Then the bullet hole image was intensified by image morphological operation. Finally, through differential operation of the edge features in two plots, the positions of bullet holes were obtained. Bullet holes in the final position by the geometric relationship with concentric circles and polar coordinates automatically identify the number of rings. Verified that the algorithm is faster time, adaptability, judgment precision can reach more than 80%, basically meet the actual needs of sports shooting competition.

This paper proposes, designs and constructs an image recognition & voice technology based security system by highlighting the advantages of image processing technology and voice synthesis technology which are presence in the electronic market. This paper mainly approaches towards enhanced security by checking the tag image of the operator as well as recognizing voice which are validated previously by this system just using simply a web-camera or closed circuit television (CCTV) camera and a voice recording software system and gives the signal in terms of alarm, Alert Light, message via global system for mobile communication (GSM)/ general packet radio service (GPRS) to the consumer mobile number/the nearest police station’s mobile number.