A low-cost smartphone controlled sensor based on image analysis for estimating whole-plant tissue nitrogen (N) content in floriculture crops

Abstract

Monitoring and maintaining plant tissue nitrogen (N) content in the optimal range is crucial for proper growth and quality of floriculture crops. Direct laboratory measurements of tissue N content is destructive, time-consuming and expensive. The equipment costs for indirect measurement of tissue N content are high. Currently, there are no low-cost, reliable and non-destructive techniques for measuring whole-plant tissue N content in floriculture crops. The objective of this study is to develop a low-cost sensor that can instantaneously and non-destructively estimate whole-plant tissue N content in floriculture crops. The technique involves a sensor, smartphone and local computer. The low-cost sensor was built using Raspberry Pis, camera modules and light filters. A smartphone interacts with both the sensor and the local computer using wireless network connection. Using a web-interface, smartphone was connected to the sensor for capturing and transferring plant images to a cloud storage. Subsequently, the smartphone was remotely connected to the local computer with image-processing software. Images were processed to calculate a reflectance ratio (Rratio). Results indicated linear and positive relationships between the laboratory measured whole-plant N content and Rratio (r2 = 0.70) and chlorophyll concentration and Rratio (r2 = 0.71). Further, a model based on stepwise selection (N = 125) indicated that tissue N mostly affected Rratio (partial R2 = 0.66), while other elements in the plant tissue had minimal effect on the ratio. Based on these results, the developed algorithm can be used to instantaneously compute whole-plant tissue N content using the low-cost sensor.