Effect of plant canopy shape and flowers on plant count accuracy using remote sensing imagery

Abstract

Separate experiments were conducted to evaluate the effect of plant canopy shape and presence of flowers on counting accuracy of container-grown plants.  Images were taken at 12 m above the ground.  Two species of juniper (Juniperus chinensis L. ‘Sea Green’ and Juniperus horizontalis Moench ‘Plumosa Compacta’) were selected to evaluate plant shape and Coral Drift ® rose (Rosa sp. ‘Meidrifora’) was used to evaluate the presence of flowers on plant count.  Counting algorithms were trained using Feature Analyst (FA).  Total counting error, false positives and unidentified plants were reported. There was no difference between all variables measured when an algorithm trained with an image displaying regular or irregular plant canopy shape was applied to images displaying both plant canopy shapes even though the canopy shape of ‘Sea Green’ is less compact than ‘Plumosa Compacta’.  There was a significant difference in all variables measured between images of flowering and non-flowering plants when non-flowering ‘samples’ were used the train the counting algorithm in FA; total counting errors and unidentified plants was greater for flowering plants.  In this specific case, applying an algorithm that did not include a training set displaying flowers, resulted in a less accurate count.  Algorithms developed using FA appears to be fairly robust under these conditions.