Unlocking Advanced Automation with the Mojow Terrain Confidence SDK

At Mojow, we are pushing the boundaries of agricultural automation. We are excited to showcase our proprietary hardware and software technology, available to OEMs as a comprehensive SDK. This complete solution provides everything you need to power advanced automation on heavy machinery.

What is the Mojow Terrain Confidence SDK?

Our SDK integrates a custom-designed camera system with our proprietary AI model. As the machine moves, the camera feeds images directly into our AI, which generates high-fidelity terrain confidence maps. 

Unlike standard rough elevation mapping, our custom models produce detailed terrain confidence maps that correspond to exact elevation changes around the machine (such as a self-propelled sprayer), enriched with confidence values that reflect the model's certainty. This allows your machines to make smarter, more reliable decisions in real-time.

Key applications

With terrain confidence maps streaming seamlessly, OEMs can integrate our SDK to automate a variety of critical tasks, including:

Sprayer Boom Height Control: Automatically adjust boom height to perfectly follow terrain contours.

Obstacle Detection: Identify and avoid unexpected terrain variations or objects in the machine's path.

Anomaly Detection: Flag unusual field conditions that require operator attention.

Predictive Speed Control: Proactively adjust machine speed based on upcoming terrain features.

Field Roughness Estimation: Map the overall roughness of the field for data analytics and route optimization.

See it in action

Here is a glimpse of our proprietary model at work. The figures below display the original camera image alongside the corresponding terrain confidence maps.

Figure 1 - Trench Detection (Negative Terrain Confidence): In the camera image on the left, a deep trench runs alongside the crop rows. The terrain confidence map on the right accurately captures this feature as a region of negative confidence (shown in blue), indicating terrain that drops below the surrounding surface. This kind of detection is critical for applications like boom height control and obstacle avoidance, where a sudden drop-off could cause equipment damage or unsafe operating conditions.

Figure 2 - Rock Pile Detection at Distance (Positive Terrain Confidence): Here, a rock pile is visible on the horizon at approximately 60 meters from the machine. Despite the significant distance, the model correctly identifies it as a region of positive confidence (shown in red) on the terrain confidence map, indicating terrain that rises above the surrounding surface. This demonstrates the model's ability to detect elevated obstacles well in advance, giving the machine ample time to adjust speed, alter course, or alert the operator.

Figure 3 - Rock Pile and Ground Slope (Terrain Gradient Map): This figure showcases the model's terrain gradient map, which assigns a confidence value to each 2×2 meter cell of the field. In the camera image, a rock pile sits on a sloping section of ground to the left of the machine's path. The gradient map on the right reveals how the model decomposes this scene into discrete grid cells, each colored according to its gradient magnitude. The rock pile itself registers the highest gradient values (shown in bright yellow/pink, ~0.20–0.25), while the surrounding slope produces a gradient that tapers off gradually (shown in softer pink/purple tones, ~0.10–0.15) as the terrain transitions back to flat ground. This per-cell resolution gives OEMs a granular, actionable view of the terrain - not just where obstacles are, but how steeply the ground changes around them, enabling precise decisions for boom leveling, speed adjustment, and path planning. Notably, the map cell resolution is a configurable parameter within the SDK. While this example uses 2×2 meter cells, it can be set to any resolution down to 10×10 cm, allowing OEMs to tailor the level of detail to their specific application requirements.

Partner with us

By using the Mojow SDK, OEMs can fast-track their automation roadmap and leverage our state-of-the-art camera and AI systems. Whether you are building smart sprayers, autonomous tractors, or robotic harvesters, our technology is built to integrate easily into your machine's architecture.

Contact us today to learn how the Mojow Terrain Confidence SDK can elevate your machinery's automation capabilities.