Location: Sugarcane Field StationTitle: Prediction of morpho-physiological traits in sugarcane using aerial imagery and machine learning
|CHIRANJIBI, POUDYALA - Texas A&M University|
|SANDHU, HARDEV - University Of Florida|
|YIANNIS, AMPATZIDIS - University Of Florida|
|CALVIN-ODERO, DENNIS - University Of Florida|
|Coto Arbelo, Orlando|
|CHERRY, RONALD - University Of Florida|
|FIDELES-COSTA, LUCAS - University Of Florida|
Submitted to: Smart Agricultural Technology
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 8/4/2022
Publication Date: 8/6/2022
Citation: Chiranjibi, P., Sandhu, H., Yiannis, A., Calvin-Odero, D., Coto Arbelo, O., Cherry, R.H., Fideles-Costa, L. 2022. Prediction of morpho-physiological traits in sugarcane using aerial imagery and machine learning. Smart Agricultural Technology. 3. Article 100104. https://doi.org/10.1016/j.atech.2022.100104.
Interpretive Summary: Florida is the largest producer of sugarcane in the United States. Diseases, pest and abiotic stress cause changes in plant height, number of milliable stalks, biomass index, leaf chlorophyll content, and leaf area in sugarcane. Measurement of those morpho-physiological traits during the selection of new sugarcane genotypes is a labor-intensive and time-consuming task. Morpho-physiological data measured with drones can play a potential role improving the accuracy of variety selection, thus saving costs. The objective of this study was to determine the feasibility and application of drones to predict such traits in sugarcane. The analysis of data showed good results in predicting those traits with no more than 30% of error on each trait. This study suggested that drones can be a potential tool to predict morpho-physiological traits in sugarcane, reducing the cost of the breeding programs.
Technical Abstract: Morpho-physiological traits are efficient to differentiate resistant and susceptible genotypes affected by biotic and abiotic stress during the breeding process. Ground measurements of such traits in sugarcane are labor-intensive and time-consuming tasks. In this study, ground data and imageries obtained by drones were collected from plant cane and first ratoon (two site-years) field trials of the last stage (Stage IV) of the Canal Point sugarcane cultivar development program in Florida. Data were collected on soil plant analysis development (SPAD), leaf area index (LAI), plant height, normalized difference vegetation index (NDVI), and number of millable stalks per hectare. Aerial imageries were collected using a hyperspectral sensor, and ground data using handheld sensors and manual readings on three dates (April, July, and September) to determine the best timing in morpho-physiological trait prediction. Results showed that SPAD was predicted with higher accuracy (89%) compared to other traits. July was observed as the best time for predicting most of the morpho-physiological traits in plant cane and first ratoon. The NDVI values collected by GreenSeeker and drones were compared and it was found that the mean difference of NDVI values between the two sensing systems was low (0.09). This study suggested that drone hyperspectral imaging could be a potential tool to predict morpho-physiological traits in sugarcane.