Revision history [back]

Maybe you can check the function gazebo::rendering::Scene::FirstContact. It uses a camera and a pixel position (can be the camera's center) to obtain the 3D position of the first thing the ray from the camera through that pixel hits a visual. Another function is gazebo::rendering::Scene::VisualAt, which gives you the visual instead of the hit position. Note that both see the visuals, not collisions.