The spectral reflectivity of the visible light bands of a non-self-illuminating object is one of the root causes of the colour formation of an object, and its uniqueness can be used to distinguish objects. Under low illumination conditions, the ability of the visible light image sensor to capture the colour of an object is greatly reduced. However, the object still has near-infrared radiation. The near-infrared radiation of the object collected by the near-infrared image sensor is combined with the visible light information of the object to improve the colour quality of the acquired object. Since the spectral reflectivity information of an object requires professional equipment to collect, it is inconvenient to use. Aim to more convenient use of spectral information data for colour restoration under low illumination conditions. In this paper, a plant’s spectral reflectivity database in visible and near-infrared bands was proposed. Firstly, 50 plants samples such as ginkgo were collected from 400 nm to 1000 nm spectral reflectivity data by a full-bands hyperspectral sorter and processed by Minimum Noise Fraction (MNF). Secondly, according to the requirements of this database, the business process, function modules and architecture of the database are designed in detail. Finally, according to the design, use C# and SQL Server to establish the database. The establishment of this database provides some data support for colour restoration and other applications.
|