Corn emergence uniformity estimation and mapping using UAV imagery and deep learning

Assessment of corn (Zea Mays L.) emergence uniformity is important to evaluate crop yield potential. Previous studies have shown the potential unmanned aerial vehicle (UAV) imagery and deep learning (DL) models in estimating early stand count plant spacing uniformity, but few extended further field-scale mapping. Additionally, estimation date using UAV has not been achieved. This study aimed estimate map DL modeling. Corn was quantified with density, standard deviation (PSstd), mean days imaging after (DAEmean). planted at four depths (3.8, 5.1, 6.4, 7.6 cm). A system equipped a red, green, blue (RGB) camera used acquire images 10 m above ground level 32 planting (20 V2-V4 growth stage). pre-trained convolutional neural network, ResNet18, three parameters. Results showed accuracies testing dataset for PSstd, DAEmean were 0.97, 0.73, 0.95, respectively. The developed method had higher accuracy lower root-mean-square-error density DAEmean, indicating better performance than previous studies. case conducted assess different field scale. From this, maps produced. that average decreased PSstd increased increasing depths, deeper caused less later spatial this field. These could be future agronomic on temporal making decisions commercial production.