Optimization of an Optoelectronics-Based Plant Real-Time Discrimination Sensor for Precision Agriculture

Abstract

We present a novel design for an optoelectronics- based real-time plant discrimination sensor, which enables a high signal-to-noise ratio, and hence a low false positive rate, to be attained. The sensor employs graded thin film coatings and a high-speed photodetector array in conjunction with a solar spectral filter and 3-wavelength sequentially driven laser module to generate multiple uniformly distributed laser spots of uniform optical power distribution. Plant discrimination is achieved through the measurements of the slope of the spectral response at discrete wavelengths as well as the Normalised Difference Vegetation Index (NDVI). The maximisation of the optical signal to noise ratio achieved by employing a custom designed optical band-pass filter minimises the daylight background noise saturating the photodetectors, thus enabling plant identification to be carried out with a background light intensity as high as 10,000 lux without implementing any shade construction.