Wheat ear counting in-field conditions: high throughput and low-cost approach using RGB images

Table 1 Percentage of success of the automatic counting at the original RGB resolution, greyscale and the resized imagery validation results

Trial, date of sampling		Original RGB	Greyscale	×1/2	×1/4	×1/8	×1/16	×1/32
Aranjuez, May 12	μ	92.39%	88.52%	92.14%	91.6%	88.98%	81.10%	62.94%
Irrigated	σ	6.23	9.90	5.89	6.04	7.06	8.75	7.51
72 images	r	0.79	0.73	0.78	0.78	0.76	0.71	0.64
Aranjuez. May 12	μ	91.06%	90.78%	90.30%	89.25%	85.50%	77.41%	60.12%
Irrigated	σ	6.37	8.99	6.29	6.79	7.48	8.66	6.89
24 images	r	0.78	0.76	0.77	0.79	0.77	0.76	0.74
Aranjuez. May 12	μ	91.70%	93.01%	91.15%	89.41%	84.92%	76.59%	59.59%
Rainfed	σ	6.96	4.57	7.79	8.7	9.37	8.82	7.60
24 images	r	0.72	0.80	0.69	0.67	0.65	0.62	0.47
Valladolid. June 9	μ	89.79%	80.56%	89.22%	87.67%	82.47%	72.47%	50.97%
Irrigated	σ	10.14	12.19	10.52	11.07	12.47	15.32	14.09
24 images	r	0.87	0.87	0.87	0.86	0.824	0.80	0.73
Valladolid. June 9	μ	31.86%	65.36%	31.12%	29.64%	27.01%	22.65%	14.02%
Rainfed	σ	7.54	11.53	7.38	7.02	6.51	5.8	4.32
24 images	r	0.42	0.35	0.39	0.38	0.39	0.34	0.34

Different sites and phenological stages across the set of 24 durum wheat varieties were assayed. Values presented are the means of percentage of success (μ), standard deviation (σ) and Pearson correlation coefficient (r)

ISSN: 1746-4811