/ / / Tab 1

Research Article

Global Wheat Head Detection (GWHD) Dataset: A Large and Diverse Dataset of High-Resolution RGB-Labelled Images to Develop and Benchmark Wheat Head Detection Methods

Table 1

Characteristics of the experiments used to acquire images for GWHD dataset.


Sub-dataset name	Institution	Country	Lat (°)	Long (°)	Year	No. of dates	Targeted stages	Row spacing (cm)	Sowing density (seeds·m²)	No. of genotypes

UTokyo_1	NARO & UTokyo	Japan	36.0 N	140.0 E	2018	3	Postflowering	15	186	66
UTokyo_2	NARO & UTokyo	Japan	42.8 N	143.0 E	2016	6	Flowering	12.5	200	1
Arvalis_1	Arvalis	France	43.7 N	5.8 E	2017	3	Postflowering-ripening	17.5	300	20
Arvalis_2	Arvalis	France	43.7 N	5.8 E	2019	1	Postflowering	17.5	300	20
Arvalis_3	Arvalis	France	49.7 N	3.0 E	2019	3	Postflowering-ripening	17.5	300	4
INRAE_1	INRAE	France	43.5 N	1.5 E	2019	1	Postflowering	16	300	7
USask_1	University of Saskatchewan	Canada	52.1 N	106 W	2019	1	n.a	30.5	250	16
RRes_1	Rothamsted research	UK	51.8 N	0.36 W	2016	1	n.a	n.a	350	6
ETHZ_1	ETHZ	Switzerland	47.4 N	8.6 E	2018	1	n.a	12.5	400	354
NAU_1	Nanjing Agric. University	China	31.6 N	119.4 E	2018	1	Flowering	20	300 or 450	5
UQ_1	UQueensland	Australia	27.5 S	152.3 E	2016	1	Flowering -ripening	22	150	8

Images were checked carefully to ensure that heads have fully developed and flowered.