2018 ASHS Annual Conference

Tissue N content should be maintained at optimal level to achieve the maximum growth and good quality in floriculture crops. Laboratory analysis is the only direct method available to growers for accurate measurement of whole-plant N status. However, this method is expensive and time-consuming. Other indirect methods (chlorophyll meters and normalized difference vegetation index (NDVI) sensors), can only measure small leaf sections or contain background signal when measured on groups of plants, especially small plants. There is an urgent need to develop alternate techniques for easy, rapid, accurate and inexpensive measurement of tissue N content. Tissue N affects chlorophyll content of leaves, thereby red light absorption/reflectance in plants. Thus, measuring changes in red light absorption or reflectance can be used to indirectly measure changes in tissue N content. The purpose of the current study is to test the efficacy of camera based image analysis technique as an alternative to indirectly measure whole-plant N content. An experiment was setup in greenhouse maintained at 26/20 °C (day/night) temperature and daily light integral (DLI) of 10-20 mol·m^-2·d^-1. Petunia (Petunia × hybrida L. ‘Hurrah Peppermint Stick mix’) and poinsettia (Euphorbia pulcherrima Willd. ex Klotzsch ‘Maren’) plants were grown under five different fertilizer treatments with electrical conductivity (EC) of 0.75, 1.5, 2.0, 2.5 and 3.5 dS·m¯¹ (74, 148, 198, 248, and 346 mg·L^-1 N) to generate a range of tissue N levels. A TopView imaging station with multi-spectral camera was used for capturing grayscale images at red (r, 630nm) and near infrared (nir, 870 nm) wavebands. Each pixels on a grayscale image contain information on the extent of reflected light from plants. The grayscale images were analyzed using MultiSpec V2.0 image processing software to obtain mean gray or reflectance value (R625 and R870) of plant pixels in the image, from which an image derived reflectance ratio (R₆₃₀/R₈₇₀) was calculated. Tissue N content was measured at harvest using laboratory analysis. Results showed that tissue N content increased linearly with increasing fertilizer EC. A statistically significant inverse relationship was observed between tissue N content and R₆₂₅/R₈₇₀ in both species indicating that the ratio can be used to indirectly estimate whole-plant N content in poinsettia and petunia. We are currently developing cheap accessories for smartphones, which can aid in image capture, image processing, generation of reflectance ratios and estimation of tissue N based on algorithms for easy, rapid, accurate and inexpensive estimation of tissue N content in plants.