Volume estimation of strawberries, mushrooms, and tomatoes with a machine vision system

Figures & data

Figure 4. Roma tomato (n = 30) density: mean estimated density (imaging method) versus mean actual density (buoyant force method).

Table 1. Volume and density of salad tomatoes, roma tomatoes, mushrooms, and strawberries.

				Buoyant force^b method		Omnisurface method
Produce	n	Weight^a (g)	Instron Force^a (gf)	Volume^a (cm^3)	Density^a (g/cm^3)	Volume^a (cm^3)	Density^a (g/cm^3)	% difference^a buoyant force versus OmniSurface density
salad tomatoes	30	223.1 ± 76.6	5.8 ± 2.4	228.9 ± 67.3	0.98 ± 0.02	223.9 ± 78.8	1.00 ± 0.02	−2.3 ± 2.1
roma tomatoes	30	68.7 ± 12.7	3.4 ± 2.3	72.1 ± 12.5	0.95 ± 0.04	70.4 ± 11.8	0.97 ± 0.04	−2.3 ± 1.3
mushrooms	35	21.7 ± 13.0	18.6 ± 12.0	40.3 ± 25.0	0.55 ± 0.04	40.9 ± 25.2	0.55 ± 0.06	0.6 ± 8.6
strawberries	35	32.0 ± 11.9	3.1 ± 1.5	35.1 ± 13.2	0.91 ± 0.02	36.1 ± 14.9	0.90 ± 0.04	1.5 ± 5.5

^amean ± standard deviation (SD)

^bbuoyant force = weight + Instron force measurement

Figure 1. Mushroom (n = 35) measurement accuracy comparison: mean estimated volume (imaging method) versus mean actual volume (buoyant force method).

Figure 2. Strawberry (n = 35) measurement accuracy comparison: mean estimated volume (imaging method) versus mean actual volume (buoyant force method).

Figure 3. Mushroom (n = 35) density: mean estimated density (imaging method) versus mean actual density (buoyant force method).

Table 2. Fruit volume estimation from weight measurement.

Produce	n	Weight^a (g)	Weight range (g)	Predictive Equation for Volume (OmniSurface) from Weight	R^2
salad tomatoes	30	223.1 ± 76.6	106.4 – 422.0	Volume (cm^3) = −5.0 + 1.03 Weight (g)	0.99
roma tomatoes	30	68.7 ± 12.7	48.0 – 99.4	Volume (cm^3) = 7.3 + 0.92 Weight (g)	0.99
mushrooms	35	21.7 ± 13.0	7.7 – 49.6	Volume (cm^3) = −0.64 + 1.91 Weight (g)	0.92
strawberries	35	32.0 ± 11.9	13.1 – 60.8	Volume (cm^3) = −5.21 + 1.27 Weight (g)	0.96

^amean ± (SD)