Snap: Chlorophyll Concentration Calculator Using RAW Images of Leaves

摘要

Determination of nitrogen content, a major component of chlorophyll in plant leaves, is a demand. In this paper we propose Snap, a handheld optical sensor device that uses RAW images of leaves for determining their chlorophyll content. By skipping the pre-processing pipeline at image signal processor, RAW images help to generate a relatively linear relationship between the incident illumination and the calculated pixel value. Pixel values are used to calculate spectral parameters, which are known for their correlation with the chlorophyll concentration. Illumination on the camera lens is controlled by specific geometrical constraints, which are derived using the diffusion properties of the illumination sources. The apparatus banishes influence of external illumination. The approach, by integration with a mobile app, caters to spatial variation of nutrients in soil. A pilot implementation involving 50 cotton plants was conducted for comparing the spectral-parameter value calculated by Snap and SPAD 502 plus Chlorophyll meter. A correlation factor of 0.908 was observed, thereby proving significant similarity but at a tremendously lower cost.