Co-chairs : Takuji Kiura
J. Adinarayana
Ye-Nu Wan
Pisuth Paiboonrat
Members :
Objectives : In Workshop, ICT application in Agriculture will be reported by participants from several APAN regions including Thailand, India, Sri Lanka, and Japan. Some of them focused on Impact Estimation Climate Change to Main Crops in APAN regions. Some report will be related to Sensor Network Application.
In the WG meeting, we will discuss about how to organize new joint project sharing Crop Model Service within APAN regions and the progress of AGWG showcase will be reported.
Target Audience : Open for all, especially the IoT researchers and SN users on many fields such as agriculture, military, disaster, and bio-inspired system.
Expected Number of Participants : 20 - 30
Agenda : 26-08-2012

Session Chairs: Masayuki Hirafuji, Takuji Kiura, J. Adinarayana
Speakers:

  • Masayuki Hirafuji, Personal Fabrication for Agriculture - (Presentation)
  • J. Adinarayana (IITB, India), Information, Communication and Dissemination Technology (ICDT) activities at CSRE, IITB - (Presentation)
  • Kei Tanaka (NARO Japan), A Framework for Applying GPU Parallel Computing to an Agricultural Model

    Abstract:
    It is necessary to execute a crop growth model tens of thousands of times to generate the one year data for the tool for predicting the possibility of rice, developed by the DIAS project, at each of the fifteen thousand points (throughout the world in one degree grid).When this calculation was executed with PC equipped with Core i7-860, seven days were required. Execution time was shortened by applying multithreading in two days. However, a further shortening of the execution time is necessary to increase the number of target crops, and to increase the number of points.
    Recently, General-Purpose Computing on Graphics Processing Units(GPGPU) attracts attention in High Performance Computing (HPC) field. The field where GPU is suitable is Single Instruction Multiple Data(SIMD) operation that executes the same processing at a time for different data by one instruction. It is thought that there is an advantage of applying GPGPU to execute a large amount of same calculations with different meteorological data and parameter values for a crop growth model.
    Aparapi, which is a Java library to use Open Computing Language(OpenCL), was adopted from the following reasons: (1) it is possible to develop in Java to use crop models implemented in Java; (2) it can be speed-up by multithreading execution with CPU when GPU is not installed in PC; and (3) it is not limited to GPU of a specific manufacturer. The DVI calculation part of SIMRIW was executed as test program, and the execution time was measured by three types of execution methods, normal, multithreading, and GPU.
    Because the GPU execution had the overhead of the calculation beginning, it was slower than normal execution up to tens of thousands of times, and slower than multithreading execution up to several hundred thousand times. However, the execution time has not increased to one hundred thousand executions. Therefore, it became the execution speed of tens of times of normal execution in case of several hundred thousand times execution. It was clarified that GPU was effective for the crop growth model calculation with the fairly large frequency. - (Presentation)

  • Masayuki Hirafuji (NARO Japan), Toward Global Agricultural Cloud - (Presentation)
  • Kiyoshi Honda (Chubu Univ., Japan), Kumpee Teeravech (Asian Institute of Technology), Amor V.M. Ines (Columbia University) and Rassarin Chinnachodteeranun (ListenField Co., Ltd.) Tomorrow's Rice Ver. 1: Rice yield simulation and prediction by DSSAT on Web - (Presentation)
  • Pisuth Paiboonrat (NECTEC Thailand), Information as a Service for Tailor Made Fertilizer in Paddy Field.

28-08-2012 Group Meeting

  • Shin-ichi Sobue and Kei Oyoshi (JAXA Japan), Asia Rice Crop Monitoring using space technology (remote)

    Abstract:
    Asian countries are responsible for approximately 90% of the world's rice production and consumption. As a result, rice is the most significant cereal crop in Asia. The Japan Aerospace Exploration Agency (JAXA) worked with the Thai Geo-Informatics and Space Technology Development Agency (GISTDA) to develop a prototype system designed to provide crop anchorage and yield estimation. The system uses space-based radar and other earth observation satellite-based data along with ground observation data and a crop growth model for rain-fed rice in Thailand. This paper presents an overview of this prototype system and provides early validation results. - (Presentation)

  • Topic 1 (Country/Regional Short Report)
  • Topic 2 (AG-WG Progress Report) Takuji Kiura, J. Adinarayana, Pisuth Paiboonrat, Ye-Nu Wan (APAN AG-WG co-chairs), APAN AG-WG Show-case: A System of Rice Field Sensor Network Systems in Asia. - (Presentation)
  • Topic 3 (Project Proposal if any)
  • Topic 4 (Next APAN Meeting), Session Proposal from AG-WG to TIP2013/APAN35th

Remarks : Video Conference System is required.

© Copyright 2009-2012 APAN | | Last updated: 22 Jun 2012