Buhodra Ingeniería, vision technologies to identify road hazards

The Buhodra Engineering project

The project developed by Buhodra Ingeniería is titled “Vision technologies applied to the identification of elements in roads” and is carried out within the framework of AsDIH, together with our partner Idonial.

This initiative was born with the aim of exploring the use of advanced vision, identification, and image processing technologies to identify and locate elements in roads that could be implemented in BIM models. The elements to be located focused on traffic signs, guardrails and road markings.

What are the objectives?

In this service, the exploration of several technologies was proposed.

Identification and positioning of signs from point cloud images. Visualization of road markings and guardrails was explored, although it was limited by the inherent characteristics of the input data format. The client provided geo referenced image data and, from these, we sought to identify, through advanced vision techniques, both the elements to be detected and their position relative to the geo positioning of the starting image. In this way, they were identified as output data and their position regarding the image positioning label.

Identification and positioning of road markings from orthophotos. The client provided the geo referenced images. From these images, we sought to identify, by advanced vision techniques, both the elements to be detected and their position relative to the geo positioning of the starting image. In this way, the type of element identified and its position regarding the image positioning label are identified as output data. Due to the characteristics of the type of image (zenithal view), it will focus on the identification of road markings.

Identification and positioning of guardrails, combining orthophotos and point cloud images. To address the limitations of each of the separate imaging technologies, both solutions were combined to verify if the technology yields better results. The client provided image data extracted from the geo referenced point clouds it manages, as well as orthophotos. From both images, advanced vision techniques were used to identify both the elements to be identified and their position relative to the geo positioning of the starting image, so that the type of element identified and its position regarding the image positioning label were identified as output data.

What technologies were used?

Robotics, artificial vision, sensor technology and data analysis.

What was achieved with this project?

The identification of traffic signs and road markings was achieved through the use of artificial intelligence, applied to orthophotos and images obtained from point clouds.