EyeVi feature extraction

---

Mobile Mapping Equipment 

EyeVi: DataCapture 

EyeVi: DataFlow 

MMS products 

EyeVi: WebApplication ​

EyeVi feature extraction

EyeVi: 2-pager

---

​​The EyeVi system can detect various elements on road surfaces or in the vicinity of roads. The extraction can be done from panoramic view, orthophoto view and LIDAR point cloud. In accordance with the clients’ needs, the EyeVi feature extraction can detect and classify e.g. posts, traffic signs, street lights, curbstones, ditches by the road, markings on road surfaces, safety islands, road surface defects, road attributes such as width, type of pavement, number of lanes, bus stops, greenery, road surface models, ground surface models, buildings etc.

---

Orthophoto view based feature extraction

​The EyeVi solution is easy to use and a productive tool to verify, update and make changes to road or other object databases. All data can be connected with a road ID, location and road linear reference. The following data can be managed:

• Road attributes – pavement type, road width, number of lanes, angle of curves, road incline and decline.
  
  
• Road objects:
  ​
  • Curbstone – location (line), condition
  • Sidewalk/pedestrian road – location (line), pavement, width, condition, separation from the road (type and width)
  • Road signs – location (point), type, condition, impact direction, text on the sign.
  • Traffic lights – location (point), direction of lights
  • Streetlights – location (point), type of post, number of lamps.
  • Sewage manholes – location (point), type, diameter.
    
    
• Pavement condition and defects:

---

Pothole-location (point)

---

---

Transverse crack-location (line), length

---

---

Weathering-location (polygon), area

---

---

Joint reflection crack- location (line), length

---

Linear crack-location (line), length

---

---

Crack network-location (polygon), area

---

---

Patched road- location (polygon), area

---

---

Edge defect- location (line), length

---

---

​It is possible to visualize road defects digitized in orthophoto view on panoramic images.

---

Panoramic photo based feature extraction

---

​It is possible to identify many object categories from panoramic photos. Here are some examples:

• Traffic signs

---

• Other man-made structures (e.g. street posts)

---

• Borders (footprints) of buildings

---

LIDAR point cloud based feature extraction

​Feature extraction from 3D point cloud is the most accurate. It enables to get coordinates of an object much more precisely than from photo images. It also gives the exact geometry of objects.

• Detection of objects on, above and near a street

---

• A 3D-layer of street posts

---

HD mapping

​​HD maps give exact images of roads and their surroundings. In general, one might describe it as a topographical plan with a special focus on road objects, traffic signs and traffic management tools, lanes, road surface markings etc. HD maps are especially useful for developers of autonomous vehicles, geodesists, road maintenance, traffic control and local governments. For the autonomous vehicle industry, the locations and restrictions of lanes are very important.
​

• Extraction of features

---

• Vectorizing the features

---

• Production of vectorized navigational information