Fractional cover data retrieval using eCognition

For the overall calibration and validation of the AquaCrop model decent information on
the development of the canopy cover is crucial. Throughout the growing season overhead
canopy pictures were taken. Different image treatment programs have been evaluated on their
capacity to retrieve the fractional cover data from these images. The findings are presented in
this technical note. The objectives (and/or constraints) were manifold:
- determine for each image the percentage of canopy cover of the monitored crop (maize
or cabbage);
- in case of mixed-cropping, calculate a separate canopy cover for each crop;
- weeds are to be excluded;
- mask all non-desired items (camera operator, field borders, …).