29
the australian farmer
not one Agtech could do it all, and that it was the application that best dictated the best tech and analysis methodology to use. Professor Andrew Robson is Director of the Applied Agricul- tural Remote Sensing Centre at University of New England. For more than a decade, he was also senior research scientist (Agricultural remote sensing) in the Queensland Government’s Department of Agriculture and Fisheries.
be a reliable predictor for tree crop yield at the block and farm level (this methodology is now being deployed across seven countries); thermal imagery was the most accurate for identifying specific diseases such as Phyto- phthora; Specific wavelengths from hyperspectral imagery that were found to correlate with specific constraints were gen- erally not transferable to other growing locations and seasons due to the influence of other growing constraints. The moral of the story is that
for yield forecasting i.e. time of day, flight direction, image overlap, speed etc. Then it was found that no sensors could ac- curately count all the fruit on avocado and citrus trees due to fruit growing within the canopy, in short if you can’t see it then neither can optical sensors. Ground based sensors were able to count fruit on mango as the fruit hang below the can- opy; historic imagery identify- ing seasonal growth patterns and their subsequent correla- tion to annual yield proved to
Back to Contents
Powered by FlippingBook