Creating graphs
Using built-in visualisation tools to explore detection results
The second built-in visualisation tool allows to visualize the time series of the vegetation index with the associated model, the anomaly detection threshold and the associated detections, for a given pixel.
Creating graphs of time series evolution at pixel level
This tool creates figures including :
- the vegetation index value for each SENTINEL-2 acquisition
- the corresponding harmonic model
- the threshold used for anomaly detection
- the period used for training.
This allows illustration of the dynamic related to anomaly detection for pixels of interest. The following illustration displays time series for a healthy pixel, and for a pixel experiencing bark beetle outbreak :
| Healthy pixel | Attacked pixel |
|---|---|
 |
 |
This process can be performed using a shapefile containing points if parameters shape_path and name_column are used, or with parameters x and y to plot a single pixel chosen from coordinates in the SENTINEL-2 data CRS. If none of those parameters are used, the program will prompt the user in a loop to enter coordinates.
In this example, we used a shapefile provided in the fordead_data repository.
Comprehensive documentation can be found here.
Running this step using a script
Run the following instructions to perform this processing step:
from fordead.visualisation.vi_series_visualisation import vi_series_visualisation
vi_series_visualisation(data_directory = data_directory,
shape_path = "<MyWorkingDirectory>/vector/points_for_graphs.shp",
name_column = "id",
ymin = 0,
ymax = 2,
chunks = 1280)
Running this step from the command prompt
This processing step can also be performed from a terminal:
fordead graph_series -o <output directory> --shape_path <MyWorkingDirectory>/vector/points_for_graphs.shp --name_column id --ymin 0 --ymax 2 --chunks 1280
Outputs
The plots are saved as .png files in data_directory/TimeSeries. One file is stored for each point with the value in the column name_column as file name. The y axis limits are set using ymin and ymax parameters.
NOTE : The chunks parameter is not necessary in this case, since the process is applied on limited surfaces. However, it is needed to reduce computation time in large datasets.