Peppers: A New Crop Implemented

Following successful deployments in tomato and cucumber production, FRAVEBOT has now expanded its multi-crop capabilities to include bell peppers.

During a one-month pilot with an innovative Dutch producer, we validated the Scout’s performance in a new greenhouse environment and collected valuable data for AI model development.

Implementation of the New Crop

The first deployment of the FRAVEBOT Scout in pepper production took place in September.

The installation process — including connectivity setup and greenhouse mapping — was completed within a single day, identical to other high-wire crops.

In the following weeks, our team focused on image data collection and model training to recognize the key structures of the pepper plant — leaves, stems, and fruits. Within three weeks, the system achieved full autonomous navigation and began collecting datasets for pest and production monitoring.

Main Focus of the Pilot

The pilot project focused on three key goals:

1. Autonomous Navigation – ensure smooth operation and row-to-row movement without manual intervention.

2. Pest Detection – collect and label image data for Nezara viridula, a major pest that increases scouting and labor costs in pepper production.

3. System Reliability – confirm consistent operation and resistance to challenging greenhouse conditions.
   

   

   
   

Insights from the Trial

The Nezara viridula pest is particularly difficult to detect due to its behavior — the insects are most visible during sunny periods and tend to hide deep in the crop at night. Despite this, early results confirmed that the FRAVEBOT Scout can successfully capture high-quality visual data to train detection models.

In addition to pest recognition, the trial also included pepper fruit detection and counting, addressing one of the key challenges in precision production monitoring for pepper growers.

Proof of Concept for Future Detection Models

The successful detection of Nezara viridula on both leaves and fruits serves as a proof of concept for FRAVEBOT’s pest-recognition technology.

This approach will be further adapted for other major greenhouse pests — specifically whiteflies in tomato and aphids in cucumber — to expand autonomous crop protection monitoring across multiple crops.