Achieve high consistency & accuracy of green coffee quality control at farms, purchase points, Q Labs & roasteries.
AGRIVERO.AI QUALITY CONTROL
AGRIVERO.AI SOLUTION ENABLES
Consistent & accurate quality control along the entire supply chain.
Consistent & accurate quality control along the entire supply chain.
Manual analysis of green coffee is a very inaccurate and time-consuming process. Accuracy of results depends on the qualification of green coffee graders and their experience, how tired they are.
There is a need for the objectivity of the quality control in the coffee supply chains in order to make them traceable and fair.
There is a need for the objectivity of the quality control in the coffee supply chains in order to make them traceable and fair.
QUALITY CONTROL REDESIGNED
Agrivero.ai Solution
Agrivero.ai solution is designed to support the current green coffee grading process at various locations: farms, purchase points, Q labs, and roasteries.
The solution consists of:
The solution consists of:
- Portable lightbox and table
- Agrivero.ai Mobile App
- Agrivero.ai Web App
Accuracy in the analysis of defects, screen sizes, and colors.
97%
3x
Faster quality checks in comparison to green coffee grading.
100%
Consistency of results across the locations.
How it works
Load a sample on the table place in the lightbox
Spreading a sample on the table and placing it in the lightbox allows producing images of consistent quality. The table is optimized for 70 g samples and seamless flipping.
Take pictures with Agrivero.ai Mobile App
Using Agrivero.ai Mobile App, take high-resolution pictures of a sample from both sides under visible light, in future versions also under UV and NIR light.
AI algorithms will do the rest
Each bean is segmented and separately analyzed. Bean colors (RGB values and intensity) are measured, bean outline shapes are drawn and bean screen size distribution is calculated. Defective beans are classified according to SCA Standard.
View, share and export results
Results are displayed within minutes in Agrivero.ai Web App. They are easily sharable via email along the supply chain and can be exported in .pdf and .csv files.
Lightbox & Sample Trays
Lightbox size: 30x30x30 cm
Sample trays: 30x30x2 cm (2 parts)
Weight: 250 g
Power supply: 5V DC
Portability: portable and foldable
Sample trays: 30x30x2 cm (2 parts)
Weight: 250 g
Power supply: 5V DC
Portability: portable and foldable
Mobile Application
Lightweight Mobile app
Download via Play Store
Operating system: Android >=Version 9
Camera: Back camera >=12 MP
Download via Play Store
Operating system: Android >=Version 9
Camera: Back camera >=12 MP
THE POWER BEHIND
AI algorithms learn the bean appearance in all their variation.
AI algorithms learn the bean appearance in all their variation.
Our AI Algorithms are trained on thousands of samples, graded by certified and experienced green coffee graders.
AI Algorithms learn all variations of beans and are able to easily analyze every variety and origin of Washed Arabica.
AI Algorithms learn all variations of beans and are able to easily analyze every variety and origin of Washed Arabica.
Defects
We detect and classify Primary and Secondary defects, count defected beans and "full bean equivalent" according to SCA guideline.
In order to classify 16 classes of defects correctly, AI Algorithms analyze colors, shapes and sizes of defected areas.
In order to classify 16 classes of defects correctly, AI Algorithms analyze colors, shapes and sizes of defected areas.
Grading
We detect bean size, screen size distribution and variation between screens.
In order to detect screen sizes correctly, algorithms calculate the area of beans and analyze their shapes based on their detected outline, minimum and maximum axis lengths and solidity.
In order to detect screen sizes correctly, algorithms calculate the area of beans and analyze their shapes based on their detected outline, minimum and maximum axis lengths and solidity.
Color
We detect bean size, % per screen and variation between screens.
In order to detect the color of the sample correctly,
AI algorithms access the color of every "good" beans, analyze variations and detect an average color per sample.
In order to detect the color of the sample correctly,
AI algorithms access the color of every "good" beans, analyze variations and detect an average color per sample.
TAKE THE FIRST STEP
Want to benefit from data that will make you successful in tomorrow's coffee environment?
Want to benefit from data that will make you successful in tomorrow's coffee environment?
Here is how we can help.
We share with you best practices of how to benefit from data, help you to access your current state, or brainstorm how your coffee business can benefit from data.
COMPANY
Agrivero UG
Derfflingerstr. 29
10785 Berlin
Germany
hello@agrivero.ai
+49 176 72665607
Derfflingerstr. 29
10785 Berlin
Germany
hello@agrivero.ai
+49 176 72665607
Subscribe TO OUR UPDATES
© 2021 Agrivero UG
