http://chineseinput.net/에서 pinyin(병음)방식으로 중국어를 변환할 수 있습니다.
변환된 중국어를 복사하여 사용하시면 됩니다.
Development of image processing techniques for the detection of occluded fruits
( Tran Quoc Huy ),( Byoung-kwan Cho ) 한국농업기계학회 2016 한국농업기계학회 학술발표논문집 Vol.21 No.2
Fruits detection is an important task for the estimation of fruits yield before harvest. In this study image processing techniques were explored to enhance the detection accuracy for occluded fruits in natural environment. Several image processing techniques, such as OTSU dynamic threshold segmentation method, characterizing color features in different color spaces, image noise removal methods and edge detection methods were combined. The overlapped fruits and severely occluded fruits by the branches and leaves were separated from the main group to follow different image processing procedures. Respective image processing procedures for the three conditions of non-occluded, overlapped and severely occluded fruits were developed.
임베디드 시스템을 위한 멀티태스킹 딥러닝 학습 기반 경량화 성별/연령별 추정
( Huy-tran Quoc Bao ),정선태 ( Chung Sun-tae ) 한국정보처리학회 2020 한국정보처리학회 학술대회논문집 Vol.27 No.1
Age estimation and gender classification for human is a classic problem in computer vision. Almost research focus just only one task and the models are too heavy to run on low-cost system. In our research, we aim to apply multitasking learning to perform both task on a lightweight model which can achieve good precision on embedded system in the real time.
Tran, Ngoc-Huy,Choi, Hyeung-Sik,Kim, Joon-Young,Lee, Min-Ho Korean Society of Ocean Engineers 2012 International journal of ocean system engineering Vol.2 No.4
For the accurate estimation of the position and orientation of a UUV (unmanned underwater vehicle), a low-cost AHRS (attitude heading reference system) was developed using a low-cost IMU (inertial measurement unit) sensor which provides information on the 3D acceleration, 3D turning rate and 3D earth-magnetic field data in the object coordinate system. The main hardware system is composed of an IMU sensor (ADIS16405) and TMS320F28335, which is coded with an extended kalman filter algorithm with a 50-Hz sampling frequency. Through an experimental gimbal device, good estimation performance for the pitch, roll, and yaw angles of the developed AHRS was verified by comparing to those of a commercial AHRS called the MTi system. The experimental results are here presented and analyzed.