THE DEVELOPMENT AND INTEGRATION OF REMOTE SENSING, GIS AND DATA PROCESSING TOOLS FOR EFFECTIVE HARVEST MANAGEMENT

For many years, Mackay Sugar has used proportional re-estimating to predict the size of remaining crop to be harvested during a harvest season. This process uses a computer program to formulate actual to estimate yield trends for the calculated harvested area, and then apply that trend to the remaining crop estimate. The re-estimate is used as the base for a harvest management system. This process required considerable input from field staff to calculate the area harvested on a regular basis and a less labour intensive method to calculate harvested area needed to be developed. Satellite imagery was adopted as it offered the opportunity to capture a snap shot of the whole district instantly and subsequent analysis of the images using change detection techniques could be used to calculate area harvested. Imagery used for change detection has included Spot II and IV multispectral images plus Landsat7 ETM. Change detection methods have been developed and refined to overcome problems associated with the changing nature of the crop including severely drought-affected cane. Also, there was a need to incorporate remote sensing data (Mackay Sugar uses ER Mapper remote sensing software) with the existing geographical information system (GIS) package to form the basis of a harvest management system. This required a rewrite of the existing harvest management computer program to a Windows based system that allowed greater flexibility of data inputs to include output from the GIS system and web based data entries. Modifications were also required on the re-estimating program. Results to date indicate that area detection from satellite imagery analysis is within 2% of area calculated from other means such as harvester GPS tracking and manual area calculation. This paper details the development of change detection methods, GIS tools and integration of data from remote sensing to a GIS layer. The development of the harvest management software is also detailed together with the results for seasons 2001 and 2002.