Ramie (Boehmeria nivea L.) is one of the main fiber crops in China, integrated with textile, forage, building materials and soil conservation. Germplasm resources of ramie in China is rich, accordingly it is of great significance to study the diversity of phenotypic characters. UAV near-air remote sensing can grasp farmland information in real time and play an increasingly important role in the collection of crop phenotype information. In this study, visible remote sensing images of ramie were acquired by UAV and leaf images were acquired by digital camera. Eight ramie leaf features and 15 canopy color texture features were obtained by image processing. Combined with ramie growth characters, the phenotypic characters of 63 ramie germplasm resources were studied. The results showed that the germplasm resources of ramie had rich variation and wide diversity. The variation coefficients of the 10 growth characters ranged from 10.46% to 47.63%, and the leaf shape, petiole color and root shape varied greatly. The diversity of plant number, plant height and stem diameter was strong. The variation coefficients of leaf shape, canopy color and texture characters were distributed between 1.56% and 30.74%, and the diversity index was distributed between 0.98 and 1.50. Among canopy color characters, RGBVI value variation, saturation (s) genetic diversity index and high variation coefficient can be estimated to be the largest, which can better reflect the difference of different kinds of materials. The entropy (Hom) variation coefficient of ramie growth potential was the highest among the canopy texture characters. Seven principal components were obtained by principal component analysis of 24 characters, with a cumulative contribution rate of 90.48%, which could effectively reflect the main information of each trait. The cluster analysis of 24 phenotypic characters divided 63 experimental materials into four categories.