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 m²/m³, respectively. For kidney bean leaves, the minimum and maximum mean LAD values were found to be 0.081 and 0.239 m²/m³, 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.

中文翻译：

---

开发用户友好的自动地面成像平台，用于精确估计大田作物的植物表型

植物表型分析是量化不同作物的质量、光合作用、发育、生长和生物量生产力的科学。过去，植物表型分析采用网格计数和回归模型等方法。然而，网格计数方法被证明是劳动密集型且耗时的，而回归模型在计算叶面积方面缺乏准确性。为了应对这些挑战，开发了一种便携式自动平台，用于对田野图进行精确的地面成像。该平台由框架、RGB 摄像头、步进电机、控制板和电池组成。RGB 相机捕获图像，然后使用 MATLAB 软件进行处理。进行统计分析以比较从网格计数、回归模型和图像处理技术获得的结果。向日葵的图像处理技术与回归模型之间的相关系数 ( r ) 分别为 0.98 和 0.97，而芸豆的相关系数分别为 0.99 和 0.96。所有选定的向日葵叶子的叶面积密度 (LAD) 的最小值和最大值分别确定为 0.132 和 0.714 平方米/立方米。对于芸豆叶，最小和最大平均 LAD 值分别为 0.081 和 0.239 m2/m3。对所开发的自动系统的人体工程学方面进行了研究。与网格计数法相比，所开发的系统具有较低的生理参数，例如工作心率为99次/分钟，工作脉搏为18次/分钟，耗氧量为786 mL/min，能量消耗为11.5 kJ/min。因此，开发的自动地面成像系统将显着减少生理工作量和相关危害。因此，与其他技术相比，所开发的方法被证明是令人满意的，为确定植物表型提供了一种快速、有效且用户友好的方法。