Detecting landscape elements such as trees and hedges using various data sources like satellite and aerial images. Goal is to automate monitoring to ensure that subsidies are granted correctly to farmers who help maintain the typical Dutch landscape. We perform this SBIR project in cooperation with 52impact for the Ministry of Agriculture, Nature and Food Quality and the Netherlands Space Office.


Improving ZIVVER’s classification system that determines whether emails contain confidential information or not. This system aids the users of ZIVVER’s secure emailing service in their decision whether to encrypt and securely send their email. The challenge is to train a model online, without having direct access to the emails because of their sensitive content.


Zero Food Waste is developing a product which monitors the food waste of restaurants, enabling them to shop in a smart way. Using a scale and camera integrated around the trash can the amounts and types of food being thrown away are detected. We build the computer vision algorithm to automatically process the camera images: detecting the food, determining its size and classifying the type of food.


Predicting the quality of provincial roads in the province of Zuid-Holland using existing historical measurements. The goal is to enable the asset coordinator to take the road quality better into account in the planning of major maintenance, ensuring outstanding road quality at lower cost.


Detecting assets around rail tracks on aerial images using computer vision techniques. These aerial images are obtained from a geographic information system (GIS). In the end, the locations of the assets are stored back into GIS for easy accessibility. In the current project, relay boxes are localised.


Detecting crops from satellite images to check compliance of farmers with their lease contracts with the government. A crop rotation is prescribed in these contracts to avoid soil depletion and guarantee food supply for the generations to come.

[wp-video-popup video=""]


SmartDocuments is a software application in which the user can create documents, like letters, bills and contracts, by filling in questions. These question protocols are called templates and are tedious to make. This process is automated by performing information retrieval using Natural Language Processing (NLP).


All packages within a parcel sorting centre should have known locations. This used to be a labour-intensive process in which barcodes are manually scanned. This is avoided by tracking the packages in a camera feed using a deep, convolutional neural network.
+31 (0)6 104 220 98
Kluyverweg 1 – 12.15, 2629HS Delft, NL

Chamber of Commerce No.: 81087578
© 2018-2024 Geronimo.AI B.V.
All rights reserved