Image sensing is expected as one of powerful tools for precision agriculture. Technical trends of the image sensing are hyper-spectral, three-dimensional, and active sensing. Ordinary multi-spectral imaging is available for obtaining information on color, pigments, shape, and growth of individual plants and parts; changes in water status in plants and soils; and soil properties. The hyper-spectral image sensing capable of resolving several hundreds of spectral bands in the region from visible to short-wave infrared make it possible to provide exactly more information by analysis of continuous spectral properties in comparison with the ordinary multi-spectral analysis. Thermal infrared imaging, a passive spectral imaging method, is effective for early detection of plant stresses as well as for measurement of surface temperature of plants and soils. Image analysis of energy balance on the leaf and canopy provides information on stomatal response and evapotranspiration. Fluorescence image sensing, an active method, such as spectral analysis of steady-state laser-induced fluorescence (LIF) provides information on changes in cell walls bound by fluorophores and of bleaching of plant pigments. Analysis of chlorophyll a fluorescence induction is used for not only early detection of patchy changes in stomatal aperture and photosynthetic activity caused by biotic and abiotic stresses but also analyzing the development of the photosynthetic apparatus of attached leaves. Three-dimensional (3-D) surface imaging, such as analysis of stereo-pair images and laser scanners (scanning range finder), makes it possible to remotely estimate 3-D structure and growth of plants and canopy. In this paper, these image sensing techniques and the applications were reviewed briefly.

View full abstract

Rainfall to Satsuma Mandarin yard often causes disease and quality deterioration of the fruit. This could be prevented when farmers could have the data of the amount of rainfall exactly and promptly.
Wakayama Prefecture Public Works Department Erosion Control Division is currently measuring rainfall at 137 places in Wakayama Prefecture and it is open to public on the web site. This only exhibits real-time data and does not show past and accumulated data.
In the system we developed, rainfall data are automatically stored and processed into several forms, i.e., the amount of rainfall in an hour of the past 24 hours; in a day of the past month, or one month, and accumulated rainfall from any day in the past to the present day, which can be used conveniently for Satsuma Mandarin husbandry.
This system is assembled in the Arida Satsuma Mandarin Database (2002) (http://www.mikan.gr.jp) for users of both PC's and cellular phones, which enables farmers to use at their farms.

View full abstract

Although development of the management diagnosis and planning method has formed one domain of agricultural management research for many years in recent years the approach which does research and development in these techniques with the form of a computer system is seen. By development research of such a system it becomes important by evaluating a system to clarify the validity and subject. It is thought that the attribute of users also influences the evaluation result of a system with the performance and validity of the system itself. However the research which clarified this point positively is not seen. Therefore in this paper for the training participant of the FAPS the relevance of the questionnaire reply pattern and the participant attribute was analyzed statistically. The factors which specify the degree of comprehension of a system the degree of satisfaction evaluation and a use intention are clarified.
Cluster analysis was applied to the 125 training participant questionnaire and seven clusters were extracted based on the degree of similar of the reply pattern of eight questions (a total of 69 choices). It became clear by comparing these results with the training participant attribute for every cluster that a reply pattern changes with occupations (a research worker a extension staff etc.) special field of study (management or Crop) practical experience in application of mathematical programming such as linear programming exercise experience of FAPS etc.
By quantitative quantification analysis type II, it was thought that practical experience of a mathematical programming, specialty, and occupations are major factors of determining comprehensive evaluation. "Management" in speciality "Researcher" in occupations with practical experience of a mathematical programming have the positive influence on degree of comprehension of a system satisfaction evaluation and intention of use.

View full abstract

The work journal system with speech recognition enables workers on farms to reduce the workload and prevent miss-recording. This system is comprised of three sections, an IC recorder, a speech recognition system and a work content analyzer. In order to enhance the recognition preciosion of the speech recognition software, we examined a compensation algorithm for conversion errors after the data conversion into text form. By keeping work records in a database, a work item dictionary to compensate conversion errors, is provided. Evaluating text conformity rate, suitable work time and work frequency based on the dictionary, the most suitable work items can be selected. The accuracy of this system can be improved by accumulating work records.

View full abstract

For farm decision support and farm work optimization as well as farm product traceability, easy and accurate acquisition of farm work record is highly requested. A system that can acquire accurate work record with position information by GPS, could be one of solution to realize this requirement. In this study, we examined a system that estimates work records based on GPS data and automatically records them. The system automatically estimates work record by rule based reasoning and fuzzy based reasoning on the GPS based positioning data and time data with the database on field information, work record, machinery information, fertilizer/chemical information and crop information. Using tracking data on fieldwork, the system can also estimate field area and working efficiency. In this study, we demonstrate that the work identification from multiple candidates by fuzzy logic base estimation with positioning data, work tracing and velocity data based on GPS is fairly effective.

View full abstract