Research on the Construction Strategy of Information Model for Manchuria Style Architecture and Its Application

Rui Han,Wei Mo,Dan Shao 보안공학연구지원센터 2016 International Journal of Smart Home Vol.10 No.7

“Manchuria style architecture” is an important landmark of colonial architecture in modern China. It shows the combination of western classicism and oriental national style in specific historic context. Its sturdy structure and superb technology are of great academic significance for studies of cross-regional building spectrum. Based on the building information modeling technology, the mapping analysis and three-dimensional reconstruction of Manchuria style architecture is conducive to promoting the routine administration of life cycle information management of historic buildings. Besides, the stirps models date base of the building components established in the research process will provide a multi-dimensional and more accurate database for the study of modern architecture heritage with reinforced concrete frame structure. As research continues, the real-time detection information transmission interface and the virtual historic building situation will be constructed, which will open up new opportunities for the studies of the preventive conservation, cultural value and educational value of historic buildings in China.

Research on the Application of Building Information Model Technology in the Design of Urban Residential Buildings in Cold Region

Rui Han,Yueqiu Gao,Dan Shao 보안공학연구지원센터 2016 International Journal of Smart Home Vol.10 No.5

The design of urban residential buildings in cold region is different from that in low latitude region, or the design of general public buildings. The precise grasp of dwelling unit type and appearance style, effective control of building heating and thermal insulation efficiency and the implementation of the sustainable concept of low cost construction and environmental protection are dependent on the early scientific and rational design. And the building information model technology just acts as a platform for the optimization of the design. The study is based on the engineering projects; the planning, the design of building facade, collision detection optimization and the full participation of construction cost control in building information model provide a new train of thoughts and research reference for the residential design in cold region, in the context of the digital age, and to a certain extent, the study looks forward to the future urban residential design and construction industrialization prospect in the cold region . Building information modeling technology improves the design efficiency and quality with its precise data processing and analysis, avoids the defects in design and the material waste of construction and environment pollution, significantly enhancing the pre-processing level of construction structure parts, which is in line with the energy-saving environmental protection and rural environmental construction phil

Application of a Deep Learning-Based Instance Segmentation Model for Behavior Classification of Pigs

Ruihan Ma,Jin-Seong Park(박진성),Sung Hoon Kim(김성훈),Sang-Cheol Kim(김상철) 제어로봇시스템학회 2022 제어·로봇·시스템학회 논문지 Vol.28 No.4

In pig-breeding livestock farms, increasing yearly sow productivity and reducing piglet mortality are very important for farmers because they are directly related to profitability. To meet these requirements of the livestock industry, smart livestock technology is being developed and propagated. Recently, many studies have been conducted to obtain information on the health and physiological status of livestock by applying image processing- and artificial intelligence-based technology. In particular, because changes in pig behavior patterns can provide considerable information about their health and physiological status, this study proposes a method to classify pig behavior patterns from images of a pig kennel by applying a deep learning-based instance segmentation model. For this purpose, we first constructed a Pig Motion dataset by building a robot operation system (ROS)-based data collection system that could synchronize and collect voice and climate information along with pig motion images in time. This dataset was organized by classifying four types of pig behavior, namely, lying, standing, sitting, and eating, into motion classes. By learning this dataset using the instance segmentation model, an object is extracted from the pig’s motion characteristic information to create a model that divides and classifies it. We propose a behavioral pattern classification model that can be used to estimate the health and physiological state of pigs by classifying their behavioral patterns from streaming videos and accumulated data and statisticalizing them.

Study of a Deep Learning-Based audio recognition Technique for Pigs health identification

MA RUIHAN,김상철,김종린 제어로봇시스템학회 2022 제어로봇시스템학회 국내학술대회 논문집 Vol.2022 No.6

Pig sound recognition using deep learning-based unsupervised method

마리한 ( Ruihan Ma ),김상철 ( Sang-cheol Kim ) 한국농업기계학회 2022 한국농업기계학회 학술발표논문집 Vol.27 No.2

In pig-breeding livestock farms, sounds recognition has attracted more attention for pigs' health monitoring. Commonly, the past way is generally simple and easy manual monitoring, but manual monitoring not only has high labor costs but also the recognition rate is difficult to guarantee. The pig sound recognition contains pig sounds collection, pig sound preprocessing, sound feature extraction, and sound recognition. However, sound annotation work is time-consuming. We proposed a deep-learning-based unsupervised method to recognition pig sound. Firstly we build a ROS-based data collection system to collect the pig sounds. Then we will employ the deep-learning-based unsupervised to train collected data. Unsupervised method is effective to recognize the pig sound from original data. It can help research reduce data annotation workload.

Pig sound recognition using deep learning-based unsupervised method

마리한 ( Ruihan Ma ),김상철 ( Sang-cheol Kim ) 한국농업기계학회 2022 한국농업기계학회 학술발표논문집 Vol.27 No.2

In pig-breeding livestock farms, sounds recognition has attracted more attention for pigs' health monitoring. Commonly, the past way is generally simple and easy manual monitoring, but manual monitoring not only has high labor costs but also the recognition rate is difficult to guarantee. The pig sound recognition contains pig sounds collection, pig sound preprocessing, sound feature extraction, and sound recognition. However, sound annotation work is time-consuming. We proposed a deep-learning-based unsupervised method to recognition pig sound. Firstly we build a ROS-based data collection system to collect the pig sounds. Then we will employ the deep-learning-based unsupervised to train collected data. Unsupervised method is effective to recognize the pig sound from original data. It can help research reduce data annotation workload.

Pig Motion Recognition Using Instance Segmentation

MA RUIHAN,김상철 제어로봇시스템학회 2021 제어로봇시스템학회 각 지부별 자료집 Vol.2021 No.12

Smart pig breeding became more and more popular in the world. It can help the government and farmers reduce the cost of breeding animals and increase sow productivity (PSY) and reduce the rate of unnatural death of piglets. Meanwhile, Smart farm can reduce disease loss and ensure pork quality. In this paper, we employ the deep-learning-based instance segmentation technique to monitor the pig motion. Firstly, we create a new pig dataset. In our dataset, there are 4 pig motion classes including lying, standing, sitting, eating. Secondly, we utilize the state-of-the-art instance segmentation model called Mask-RCNN to extract the pig feature information. Thirdly, according to the result, we analyze the pig motion and give a suggestion for pig breeding.

동역학 시뮬레이션에 의한 농업용 UTV의 서스펜션 설계를 위한 스프링 계수 결정 시험

김상철,김성훈,MA RUIHAN,ZHOU HENG 제어로봇시스템학회 2022 제어로봇시스템학회 국내학술대회 논문집 Vol.2022 No.6

Efficient Traffic Engineering for 5G Core and Backhaul Networks

Gang Wang,Gang Feng,Shuang Qin,Ruihan Wen 한국통신학회 2017 Journal of communications and networks Vol.19 No.1

The next generation mobile networks (5G) should beefficient and elastic to accommodate numerous and diverse ser-vices. By explicitly assigning bandwidth to service flows, trafficengineering (TE) is effective to improve network efficiency andelasticity. Unfortunately, existing mobile network TE schemes aremostly focused on core network only, which is inadequate for effi-cient end-to-end traffic delivery in mobile networks. In this paper,we propose a TE framework that incorporates the data gateway(D-GW) selection and exploits the topology and traffic informa-tion of both core and backhaul networks for SDN-based 5G net-works. With ideal flow to D-GW association (IFDA) strategy, weformulate the TE problem as a multicommodity flow problem toachieve network load balancing. Considering the cooperation sig-nalling between D-GWs, we propose multiple BSs to one D-GWassociation (MBODA) and multiple flows to one D-GW associa-tion (MFODA) strategy, and formulate the corresponding TE prob-lems as mixed integer linear programs (MILPs) which are NP-hard. To efficiently solve the IFDA-TE problem, we design an im-proved version of fully polynomial time approximation scheme (i-FPTAS). Moreover, we propose a heuristic method and an LP re-laxation method that both use i-FPTAS to solve the MBODA-TEand MFODA-TE problems respectively. Numerical results showthat i-FPTAS achieves close-optimal solution with significantlylower computational complexity, compared with FPTAS, and theperformance of MFODA-TE is very close to that of the IFDA-TE,while there is a small performance degradation for MBODA-TE asthe cost of computational efficiency.