Two Efficient Techniques for Difficult Applications of Machine Vision in Construction Industry

Machine vision is prospectively the most powerful sensing technique for intelligent automation. However, the existing techniques and algorithms only partially satisfy the needs of construction industry, mainly because of degrading effects that are frequently present at construction sites (weather conditions, muddy, sandy or water-covered sites, partially buried objects, sources of intensive light, etc.). In this paper, we discuss principles of two methods that can significantly improve vision system performances in such conditions (in particular at visually difficult underwater or underground sites). The first method (based on gated imaging) is an efficient image-capturing technique while the second method addresses the problem of object detection in visually complex environments. Both methods have been tested using data and conditions very similar to those expected in real problems, but so far they have not been deployed. The methods can be used either separately or jointly, depending on the problem constraints and hardware availability.