paper Untitled

Good news for Winemakers! Digital Falcon jointly with The University of Melbourne has developed a new method to automatically segment crops with distinct row and canopy configurations (e.g. vineyard, fruit orchards and vegetable crops) in aerial imagery. The algorithm uses skeletonisation techniques, developed for fingerprint analysis, to detect the crop’s unique signature and seperate canopy pixels from non-canopy pixels. When evaluated on a commercial vineyard, the algorithm achieved a vine row detection accuracy of 97.1%. By simplifying the skill set required to collect and analyse remotely sensed data, we can start to realise the potential unmanned aerial systems (UAS) has in precision agriculture and environmental monitoring.

Description: Several studies have demonstrated that high-resolution visual/near-infrared (VNIR) vineyard maps acquired from unmanned aerial systems (UAS) can be used to monitor crop spatial variability and plant biophysical parameters in vineyards and orchards. The false colour orthomosaic (67 ha), shown in Figure 1, contains 14 ha of vine plots, although less than 10% of the orthomosaic’s pixels are actual vines. Manually separating vine pixels from non-vine pixels is time consuming, costly and often inaccurate.

MODSIM: Andrew from Digital Falcon will be presenting this algorithm in a scientific paper entitled ‘Automated detection and segmentation of vine rows using high resolution UAS imagery in a commercial vineyard’ at the 21st International Congress on Modelling and Simulation (MODSIM2015) to be held on the Gold Coast, Queensland from Sunday 29 November to Friday 4 December 2015.