The stems of widely cultivated grass species such as maize perform multiple architectural and physiological functions, while contributing the most to non-grain biomass for the production of fodder and biochemicals. Stem histology and composition are highly implicated in those functions and are highly dependent on the genotype, growing and climate conditions. Here we propose a method to quantify histology and to compare genotypes. Macroscopic imaging has been retained to study large samples such as whole stem sections of about 1 cm². At this scale, morphological features such as stem area, rind thickness, vascular bundle density can be obtained. Two kinds of images were acquired without section labelling. Visible images were obtained using a dedicated device equipped with darkfield illumination. Multispectral fluorescence images were acquired after UV and visible excitations using a fluorescence macroscope to detect lignin and phenolic acids thanks to their autofluorescence properties. Both types of imaging were used with the goal of highlighting variability of the tissues from the stem for a series of 14 maize genotypes. Rind, vascular bundles and parenchyma were segmented in both images. About 20 morphological descriptors were measured from the darkfield images. Autofluorescence pseudo-spectra were extracted from the multispectral images for each tissue. Huge variability between the 14 inbred lines was revealed. The most discriminant morphological descriptors were the relative amount of rind and parenchyma tissues together with the density and size of individual bundle, the stem area and the parenchyma cell diameter and distribution. Specific fluorescence signatures have been identified with a predominant tissue effect and the inbred line effect was always significant. In particular, the relative fluorescence emission after Blue excitation seemed to be similar within a genotype whatever the tissue. A higher level of lignin resulted in a higher visible-induced fluorescence in all tissues. The amount of para-coumaric acid was correlated with the UV-induced fluorescence of the rind and the parenchyma near the rind while ferulic acid was correlated with the parenchyma near the rind. Darkfield and fluorescence macrovision imaging techniques showed their potential to quantify tissue morphology and phenolic compounds in maize stem sections without any labelling. The image analysis pipeline is semi-automated and iadapted to following the impact of climate change on stems.