Development of a user-friendly automatic ground-based imaging platform for precise estimation of plant phenotypes in field crops

Plant phenotyping is the science to quantify the quality, photosynthesis, development, growth, and biomass productivity of different crop plants. In the past, plant phenotyping employed methods such as grid count and regression models. However, the grid count method proved to be labor-intensive and time-consuming, while the regression model lacked accuracy in calculating leaf area. To address these challenges, a portable automatic platform was developed for precise ground-based imaging of field plots. This platform consisted of a frame, an RGB camera, a stepper motor, a control board, and a battery. The RGB camera captured images, which were then processed using MATLAB software. Statistical analysis was performed to compare the results obtained from the grid count, regression model, and image processing techniques. The correlation coefficient (r) between the image processing technique and the regression model for sunflower was found to be 0.98 and 0.97, respectively, whereas for kidney bean it was 0.99 and 0.96, respectively. The minimum and maximum values for leaf area density (LAD) of all selected sunflower leaves were determined to be 0.132 and 0.714 m(2)/m(3), respectively. For kidney bean leaves, the minimum and maximum mean LAD values were found to be 0.081 and 0.239 m(2)/m(3), respectively. Ergonomic aspects of the developed automatic system were studied. The developed system had lower physiological parameters, such as working heart rate of 99 beats/min, work pulse of 18 beats/min, oxygen consumption of 786 mL/min, and energy consumption of 11.5 kJ/min compared to the grid count method. Thus, developed automatic ground-based imaging system would significantly reduce physiological workload and associated hazards. Therefore, the developed method proved satisfactory in comparison to other techniques, offering a quick, efficient, and user-friendly approach for determining plant phenotypes.