AGREMO ANALYSES

Plant Count & Health Monitoring

Agremo is an advanced analysis solution which uses insights extracted from aerial imagery to improve agricultural processes.

Drone-based remote sensing can be used for a variety of purposes in agriculture, from counting plants, early stress detection, and even for yield prediction. Agremo analyses are divided into 2 basic groups: plant counting and health monitoring.

Plant Counting

Plant counting analyses provide information about the exact number of plants. The margin of error is around 2%.

Agremo has developed 3 types of algorithms which are used for plant counting analyses:

  • Template matching
  • Row plant counting
  • Vegetative cover

Each algorithm is associated with a particular variety of input data and plant growing conditions. All algorithms work with both visual imagery and multispectral imagery. Some algorithms require additional inputs, such as recommended sets.

Template Matching

The purpose of the template matching algorithm is to automatically count individual plants in the field from high resolution drone imagery.
Template Matching is a high-level machine vision technique that identifies parts on an image that match a predefined template (plant). Agremo's advanced template matching algorithms have been designed to detect templates regardless of their orientation and local brightness.

The algorithm was developed and tested for perennial crops and vegetable such as bananas, almonds, mango, apples, lettuce, tomato, and many more.

To achieve the best possible results, avoid submitting images with overlaps between individual crops. Besides this, algorithms are unable to count plants during leaf-off season.

Supported resolution 2.5 cm/pixel or less
Outputs PDF report (with statistics), JPG result image, shapefile

 

Row Plant Counting

The row plant algorithm identifies plant rows, determines the gaps within each row and uses this kind of information to perform the actual plant count.

The algorithm was developed and tested for field crops (corn, sugarcane, potatoes etc.). The row plant counting algorithm is particularly useful to farmers and producers when used with high-resolution imagery of post-emergence crops in unifor rows.

The best results with this algorithm can be achieved with parallel and evenly distributed crop rows.

 

Supported resolution 5 cm/pixel or less
Other requirements Distance between plants
Outputs PDF report (with statistics), JPG result image, shapefile

 

Vegetative Cover (or Plants Cover)

The purpose of the vegetative cover algorithm is to automatically establish a percentage of vegetation (plants) within a 1 meter (0.5 m) radius of the image. Based on this, it is possible to compute the plant count per area. The result itself is calculated using three clusters: full vegetation 100%, moderate vegetation 50% and no vegetation 0%.

The algorithm was developed and tested for crops with high density and for cases when individual crops or rows cannot be recognized. The algorithm has shown to be very efficient for crops at various growth stages such as corn, wheat, canola, cotton, soybeans, sunflower, and many more.

Supported resolution 5 cm/pixel or less
Other requirements Information on the recommended set
Outputs PDF report (with statistics), JPG result image, shapefile

Plant Health Monitoring

Health monitoring analyses with drone imagery make it possible to measure plant health, identify crop stress and damages from different factors, and thus rapidly eliminate threats on the field. The main purpose of these analyses is to allow users to explore and benefit from detailed agricultural data using modern and reliable technology.

There are two types of plant health monitoring analyses:

  • General stress and specific stress
  • General stress

Plant stress snalyses identify the percentage and the exact location of areas with stress. In this context, “stress” refers to plants that have not emerged into healthy plants, areas without plants, areas with diseases, drought or other yield-limiting factors. Essentially, you get a map with healthy plants areas and all problem-causing areas of your field.

By this definition, roads, rocks, or even non-arable land can be considered stress zones as well. 

Supported resolution 5 cm/pixel or less
Outputs PDF report (with statistics), JPG result image, shapefile

Specific Stress

To put a name on these problem areas, we recommend performing can perform detailed analyses, based on the issue that has been detected (weed analyses, pest analysis, etc.).

These specific analyses identify the percentage and the exact location of problem areas caused by the according stress type, such as pest, weed or diseases.

Supported resolution 5 cm/pixel or less
Other requirements Ground truth information
Outputs PDF report (with statistics), JPG result image, shapefile

 

Plant Stress Analysis (General Stress)

Disease Analysis (Specific Stress)

Weed Analysis (Specific Stress)

The agriculture industry goes hand in hand with technological innovations. New technologies are used to drive results, while minimizing costs and time effort. One way to achieve this, is by making use of advanced data, which can be provided by drones. In doing so, complex workflows become easier and more efficient.

We know that every agricultural business is different. This is why enterprise users can build customized analyses with our IT and agriculture experts assisting them. If this sounds interesting, get in touch via sales@agremo.com and schedule a free consultation today.