Plant Count and Emergence for Trial Field

1. Description

Plant Count and Emergence is an analytics that automatically determines the plant count and emergence for rows of crops in field trials. Let's run through how it works.

2. Inputs

In order to properly run the analytic, you need to provide the below inputs and parameters:

Required Definition
Scouting Maps‍ 
  • Multispectral sensor: 
    • NDVI (Normalized Difference Vegetation Index) is calculated from the photogrammetry or the Generic Scouting Map analytics.
      • For row crops, with 6cm of GSD and a multispectral camera, the optimal corn stage will be 4 leaves, and starbud for sunflower. With enhanced resolution, earlier stages can be targeted.
      • The basic rule is: the size of GSD <= half canopy size (from above)
  • RGBsensor: 
    • From the Generic Scouting Map analytic: 
      • VARI (Visible Atmospherically Resistant Index) 
    • From the Custom scouting Map analytic: 
      • ExG (Excess Green) = 2*green-red-blue
      • TGI (Triangular Greenness Index) = green-0.39*red-0.61*blue
Inter-plants spacing Theoretical interplant spacing (within the row) 
It is mandatory to see the bare ground between two plants on the map
Microplots vector‍  Vector file containing microplot boundaries
Optional Definition
Row vectorization Vector file containing the digitized rows (if already available). 
Deliverables suffix A suffix applied to all deliverable names

3. Workflow

Step 1 - Within a survey, choose the "Plant count and emergence" analytic

Step 2 - Click on "LAUNCH".

Step 3 - When you enter the analytic page, select the Scouting Map that you will use to run the counting (1).

Step 4 - Select the Microplot vector that will be used (2)

Step 5 - Select the vectorized rows if applicable (3). Otherwise, this operation is done automatically.

Step 6 - Select "NEXT STEP".

Step 7 - On the parameters page, fill in the inter-plants spacing section, i.e. the distance between two plants within the same row. 

Note: The value is in meters.

Step 8 - Input a suffix for the deliverables (optional). For example: "_cropA" outputs "Plant count_cropA.geojson", and click on "LAUNCH PLANT COUNT AND EMERGENCE".

Tips: You can also indicate here the inter-plant spacing parameter in case you launch the analytic several times with different inter-plant spacings.

4. Results

From the layers panel, display the "Inventory" folder from the "Survey Data" section, and display the plant count or gaps layers.

  • If you display the "Gap" layer, it is better to also display the RGB map to identify where the gaps are located. Gaps are represented by small dots, and their color depends on the gap length. Refer to the "Legend" for color/distance definitions.

In addition, selecting the dot will also display the gap length and further attributes.

  • If you display the plant count layer, the microplots appear. The microplots will also be colored depending on the selected attribute and its corresponding values.

You can click on one microplot to display its plant number and other attributes.

5. Deliverables

Once the analytics is run, the deliverables can be exported in two standard formats, vector format such as "geoJSON," and "CSV".

List of attributes for each format:

Deliverable File Format Attributes
Gaps geoJSON
CSV
The "gaps" deliverable is only for plants in rows with the following attributes:
  • parent_id: unique id of a group of microplots
  • block_plot_id: six-figure number composed by block_row_id, and block_col_id
  • row_id: row id
  • gap_id: id of the gap
  • gap_length: length of the gap in meters
  • at_line_end: boolean which determines if the gap is at the end of a line
Rows with plant count geoJSON
CSV
  • parent_id: unique id of a group of microplots
  • block_plot_id: six-figure number composed by block_row_id, and block_col_id
  • row_id: row id
  • row_length: row length in meter
  • plant_count: number of plants in the row
  • gap_length: total length of gaps in the row in meters
  • plant_length: total length of plants in the row in meters
  • veg_ratio: plant_length/row_length, between 0 and 1
  • row_anomaly: indicates if an anomaly occurred during the vectorization
  • canopy_radius: canopy radius in meters
Plant count geoJSON
CSV
  • parent_id: unique id of a group of microplot
  • block_plot_id: six-figure number composed by block_row_id, and block_col_id
  • row_count: number of row in the microplot
  • plant_count: number of plants in the microplot
  • row_length: mean length of the rows in the microplot in meter
  • plant_length: total length of plants in the microplot in meters
  • gap_length: total length of gaps in the microplot in meters
  • canopy_radius: canopy radius in meters
  • veg_ratio : plant_length/row_length, between 0 and 1
  • row_anomaly: indicates if an anomaly occurred during the vectorization
  • anomaly: low_count, very_low_count, OR _empty_: informs the user if the plant count is far below the average number of plants
  • emergence: qualification of the early vigor of plants
  • theoretical_plant_count: theoretical number of plants
  • plant_count_ratio: ratio between the plants actually raised and the total number of plants sown or planted

Important information: 

Please note that the attributes "gap" and "length" that are displayed on the right side menu on the platform are not displayed in meters due to a conversion problem. In order to obtain this information in meters please refer to the "gap_length" and "row_length" attributes respectively, present in the downloadable CSV.

We are in the process of correcting this anomaly and apologize for the inconvenience caused. 

6. Quality check the results

This section allows identifying potential anomalies on your trial plots.

6.1 Check the "anomaly" attribute

In a plant count vector output, click on a plot to see if there is an anomaly for that given plot. Or extract the CSV file to see a report of the anomalies for all plots.

6.2 Check the "theoretical_plant_count" attribute

In the plant count vector output, if the values are very different from your theoretical plant count, there is probably an anomaly.  

6.3 Check the microplot coregistration

If the plot contour isn't well located around the plants, the results can be impacted. Use a scouting map (like NDVI) in the background to better identify the plant rows and make corrective actions.

The example below shows an overlay of microplots on a plant row. The plants present in this row are therefore difficult to count, which affects the results delivered.