4.2. Dynamic performance
Two dynamic tests were performed. In the first, each plant was passed over 5 times at a speed of 3 km/h. Due to differences in plant sizes, a total of 916 NDVI measurements for Anthurium, 305 for Dandelion and 560 for Sunkisses were captured. The results obtained are presented in the form of scatter plots Figs. 16–18, with two scatter plots given per test plant. The first scatter plot displays the discrimination of the dynamic measurements for the test plant based on the defined regions-of-classification. The second scatter plot illustrates the spread of the measurements by plotting those that, whilst identified as green, fall outside the defined regions-of-classification. In the operational context of the proof-of-concept system, these results emphasise the effectiveness of adopting the regions-of-classification definition for accurate green-from-green plant discrimination.
It is clear from Figs. 16–18 that the control data captured under laboratory conditions resulted in sub-optimal regions-of-classification for testing the outdoor operation of the proof-of-concept system. Furthermore, the spread of the measurements for the individual test plants illustrates the complexity arising from data variability obtained on performing dynamic measurements outdoors, under real-world conditions.