A method of automated detection of negative obstacles (potholes, ditches, and the like) ahead of ground vehicles at night involves processing of imagery from thermal-infrared cameras aimed at the terrain ahead of the vehicles. The method is being developed as part of an overall obstacle-avoidance scheme for autonomous and semi-autonomous off-road robotic vehicles. The method could also be applied to help human drivers of cars and trucks avoid negative obstacles — a development that may entail only modest additional cost inasmuch as some commercially available passenger cars are already equipped with infrared cameras as aids for night-time operation.

A Pothole Dug at a Construction Site looks dark in a daytime image in visible light but looks bright in an infrared image acquired at midnight.
The need for this or an alternative method arises because the geometric nature of negative obstacles makes it difficult to detect them by processing of geometric information extracted from ordinary images: As drivers of ground vehicles know from common experience, it is difficult to visually detect negative obstacles in sufficient time to avoid them, making it necessary to drive slowly enough to be able to stop or swerve within the limited safe look-ahead/stopping distance. In robotic vehicles equipped with stereoscopic machine vision or lidar systems that yield range and elevation data, obstacles are detected through analysis of those data, and essentially the same difficulty arises. The source of the difficulty is the fact that whereas the angle subtended by a positive obstacle is approximately inversely proportional to the horizontal distance, the angle subtended by a negative obstacle is approximately inversely proportional to the square of the horizontal distance.

The method involves exploitation of the fact that throughout the night, negative obstacles are usually warmer than the surrounding terrain. Therefore, in infrared terrain images acquired at night, negative obstacles usually appear brighter than the surrounding terrain(see figure). (During the day, the negative obstacles can be warmer or cooler than their surroundings, depending on sky conditions and the apparent position of the Sun.) At the present state of development, the method is embodied in a rudimentary algorithm that processes a combination of infrared imagery and range-versus-elevation data. The algorithm identifies candidate negative obstacles in the form of infrared bright spots on the terrain, then performs simple geometric tests to confirm or reject the candidate negative obstacles. The algorithm has been shown to be sufficient for initial proof-of-concept demonstrations. Further development of this or an improved algorithm will be necessary to enable reliable detection of negative obstacles under a variety of conditions.

This work was done by Arturo L. Rankinand Larry H. Matthies of Caltech for NASA’sJet Propulsion Laboratory.

The software used in this innovation is available for commercial licensing. Please contact Karina Edmonds of the CaliforniaInstitute of Technology at (626) 395-2322.Refer to NPO-40368.