Prediction of potato yield before harvest is important for making agronomic and marketing decisions. Active optical sensors (AOS) are rarely used together with other hand-held instruments for monitoring potato growth, including yield prediction. The aim of the research was to determine the relationship between manually and remotely measured potato crop characteristics throughout the growing season and yield in commercial potato fields. Objective was also to identify crop characteristics that most accurately estimated potato yield.

The research was conducted in 2018 and 2019, respectively in northern (54°31'13"N, 17°18'33"E) and central (52°4'54"N, 21°8'32"E) Poland on two (21.9 and 10.5 ha) commerial fields cropped with potato (Solanum tuberosum L.), cv. Ivory Russet and Hermes. The soil texture (ST) of the northern, irrigated field was sandy loam and the ST of the central, rainfed field was loamy sand. The crop measurement points consisted of two 1 m long ridges marked with labels located in areas of different yield potential, established using soil maps at a scale of 1:5000, and ST information and aerial images. On the northern and central field 18 and 21 sampling sites were established, respectively. The crop characteristic determined manually was crop height (cm), whereas remotely measured were: NDVI (Normalized Difference Vegetation Index) using AOS – GreenSeeker (Trimble Inc., Sunnyvale, CA, USA), LAI (Leaf Area Index) (LI-COR Biosciences, Lincoln, NE, USA), and % of green coverage of the canopy derived from images obtained with a digital camera Sony DSC-HX400V oriented in a nadir position). The images were analyzed using Image J software. The measurements on the potato field located in northern Poland were done four times at growth stages (BBCH 52; 75/79; 79/80 and 87), and on the potato field located in central Poland were performed six times at growth stages (BBCH 31; 51; 69; 75; 79 and 81). Potato was manually harvested at each sampling point at the termination of the crop.

On the fields located in northern and central Poland, the strongest correlations between crop characteristics and yield were obtained at BBCH 79/80 and at BBCH 75, respectively. In case of the northern field, crop height outperformed (R²=0.63) % of green coverage of the canopy (R²=0.46); LAI (R²=0.36) and NDVI (R²=0.34). On the central field, the opposite trend was noted: crop height did not show any correlation with potato yield but % of green coverage of the canopy was stronger correlated with yield (R²=0.65) than LAI (R²=0.57), and NDVI (R²=0.40). In conclusion, the measurement of the % of green coverage of the canopy derived from images taken with a digital camera could be a promising indirect tuber yield estimator at crop growth stages betwen BBCH 75 and 80.