On-field optical imaging data for the pre-identification and estimation of leaf deformities.

Visually nonidentifiable pathological symptoms at an early stage are a major limitation in agricultural plantations. Thickness reduction in palisade parenchyma (PP) and spongy parenchyma (SP) layers is one of the most common symptoms that occur at the early stage of leaf diseases, particularly in apple and persimmon. To visualize variations in PP and SP thickness, we used optical coherence tomography (OCT)-based imaging and analyzed the acquired datasets to determine the threshold parameters for pre-identifying and estimating persimmon and apple leaf abnormalities using an intensity-based depth profiling algorithm. The algorithm identified morphological differences between healthy, apparently-healthy, and infected leaves by applying a threshold in depth profiling to classify them. The qualitative and quantitative results revealed changes and abnormalities in leaf morphology in addition to disease incubation in both apple and persimmon leaves. These can be used to examine how initial symptoms are influenced by disease growth. Thus, these datasets confirm the significance of OCT in identifying disease symptoms nondestructively and providing a benchmark dataset to the agriculture community for future reference.