Excavation machine automation with 3D computer vision

Challenge

A road construction company was having problems with the direction of the excavation machine during tunnel projects. Making the targeting required a great time, effort, and stops that could generate risks to the project.

The project was for a construction company that used LIDAR sensors in excavations, thus generating a point cloud (Point Cloud). The project's main objective was to develop a new algorithm for detecting surface anomalies in holes in real time.

Our solution

We use LIDAR sensors on the excavator to generate a point cloud. Our main objective in the project was to develop a new algorithm that would create and perform the interpretation of this reading to make the excavation be adjusted quickly and during the process.

After several tests using computer vision tools, our team successfully developed a new algorithm based on statistics and linear algebra that identifies anomalies during excavation. In addition, we developed new software to receive new data from sensors using Ros for communication.

After that, we built a desktop software to visualize in real-time the point cloud and the defect using Python to visualize in 2D (a depth map) this same cloud.

Results obtained:

  • Fewer stops during the excavation
  • Increase in construction speed
  • Greater accuracy during excavations

Technologies:

logotipo_apollo
logotipo_apollo
logotipo_apollo
apollo_png
logotipo_apollo
logotipo_apollo
logotipo_apollo

Get in touch and request a quote for your project

Fill in your details and tell us your challenges. One of our specialists will contact you shortly to propose the ideal solution!