Estimation models of chlorophyll contents in potato leaves based on hyperspectral vegetation indices

Objective To build models for estimating chlorophyll a and chlorophyll b contents in potato, Solanum tuberosum, leaves by using hyperspectral vegetation indices, and provide a basis for scientific and nondestructive estimation of chlorophyll content in potato leaves.

Method Leaf spectral reflectance was measured for potato plants at different growth stages under different nitrogen levels by using a portable spectrometer. The hyperspectral vegetation indices were extracted from leaf spectral reflectance. The contents of chlorophyll a and chlorophyll b in potato leaves were measured, and the relationship between hyperspectral vegetation index and chlorophyll content was analyzed.

Result There were 12 vegetation indices with relatively good correlations with chlorophyll content. Modified normalized difference index(mND₇₀₅), modified simple ratio index(mSR₇₀₅), MERIS terrestrial chlorophyll index(MTCI) and modified chlorophyll absorption reflectance index(MCARI) had the highest correlations with chlorophyll contents among the 12 vegetation indices. Estimation models were established based on above-mentioned four indices. The power model based on MTCI offered the best estimation of chlorophyll a content, and the index model based on MCARI offered the best estimation of chlorophyll b content.

Conclusion The power model based on MTCI and the index model based on mND₇₀₅ can be respectively used to estimate chlorophyll a and chlorophyll b contents in potato leaves with relatively high accuracy, and indirectly monitor nitrogen deficiency in potato plants.