A convolution neural network for forest leaf chlorophyll and carotenoid estimation using hyperspectral reflectance

•A CNN was proposed for leaf Cab and Cxc content estimation using one-dimensional hyperspectral reflectance at leaf scale.•A large amount of physical data was developed to strengthen the performance of the proposed CNN.•This study expands the application of the CNNs in one-dimensional hyperspectral reflectance;•The proposed CNN has the best accuracy and generalization for leaf Cab and Cxc content estimation.