Our research is focused on developing and applying high throughput plant phenotyping tools. Our goal is to make measurements quickly, affordably and without killing the plants. Recent developments in the field of plant phenotyping have expanded image analysis to different types of cameras, such as satellites and UAV and applied new techniques in machine learning.
Current experimental challenges include the large scope and significant labor required to capture measurements at many time points, environmental treatments, and genotypes to make significant progress on biological questions. Our protocol provides a simple method for capturing and analyzing plant images with speed and flexibility without the need of expensive equipment. To begin, turn on the power strip using the switch and press the button to turn on the monitor.
Turn the switch on the single board computer or SBC power cord to On.Place the camera on a tripod or table and provide power to the camera via the power cord or batteries. Turn the camera on using the camera power button. Connect the camera to the USB port on the SBC using the provided data transfer cable and to the camera.
Next, place a piece of dark colored tape on the photography grade fabric where the pot will be placed. Now position the pot on the marked tape. For plants that grow in a plane, such as maize and sorghum, position the widest angle of the plant towards the camera.
Adjust the color card to be in line with the pot, but separated from the plant for no overlap. Click the Terminal button to open the terminal on the SBC. Enter the line of code to open the window for image capture.
Then type the name of the image manually or scan a barcode or QR code of the plant using a barcode scanner. After taking the picture, select an option to save the image locally or to a mounted server or cloud storage. Click the click here to take a picture"button to capture the image.
If local storage was chosen, open the Photos folder on the desktop to review the image. If a server or cloud storage was used, open the image in that location. After capturing all images, use the preferred method, such as USB storage, an internet browser or SSH transfer to copy all the images from the SBC to a local computer or cloud storage.
Download the required files to a local computer or server. Install plantcv using the command line or package manager. Open photo-studio-SV-notebook.
ipynb using a preferred code editor such as JupyterLab or Visual Studio Code. Run each block of code, making necessary edits based on the parameters outlined in the code to obtain a clean mask of the plant. When satisfied with the analysis of the sample image, open workflow.
py in the preferred code editor. Edit workflow. py to include any parameter changes made in the photo-studio-SV-notebook.
ipynb and save the file. Open photo-studio-SV-config. json to update the file paths.
Modify paths to input folder containing plant images and output folder where processed images and results will be stored. Open a terminal and activate plantcv with the command conda activate plantcv and press Enter on the keyboard. Change the specified line of code to point to the updated photo-studio-SV-config.
json file. In the terminal, type the modified code and press Enter on the keyboard. Ensure plantcv is activated in the terminal.
Then run the command to convert results-photo-studio. json to csv, updating the file path accordingly. Image analysis in plantcv successfully segmented maize plants from the background using a dual channel threshold.
The plantcv analysis provided 16 single value traits, including leaf area, height, width, and hue circular mean with multi value traits plotted as histograms. Leaf area, height, width, and hue circular mean explained greater than 50%of variants due to treatment, making them key metrics for downstream analysis. The plant with the largest leaf area had the highest mean hue and was a genotype B73, a well watered, heat-stressed plant.
The smallest leaf area also had the smallest mean hue and was a genotype B73, drought-stressed, heat-stressed plant. Water treatment significantly affected leaf area, height, width and hue circular mean, while temperature treatment only significantly affected height. Drought stress significantly reduced leaf area, plant height, plant width, and hue circular mean under both temperature conditions.
The reduction in mean hue due to drought was because of a shift from green to yellow pixels. Heat stress caused both yellowing and darkening of the green color. Leaf area measured via image analysis strongly correlated with plant biomass, validating the image-based phenotyping method.
