Polarization 

Specialists of our company have developed a prototype of an optical sensor that uses shooting capabilities with a polarization filter for fast and cheap 3D modelling of filmed scenes. The prototype includes a wide range of computer vision algorithms and  hardware platforms. The goal of the project is to develop 3D modelling software that uses the processing of stereo images and polarization images that work in different lighting conditions. The application of this technology has a high potential in: Archaeology and restoration, digital forensic examination, measurement of volumes in mining, forestry, agriculture and construction, analysis of oil towers and oil and gas pipelines, flaw detection of hydro constructions. 

Potential applications

Technologies of three-dimensional reconstruction find their application in the mass of science-intensive industries, such as robotics, medicine, biometrics, chemical industry, engineering, biotechnology, topography and others.
In connection with the high demand for such technologies, as well as the variety of requirements for them, we have developed algorithms for three-dimensional reconstruction, both active and passive. 

  • Analysis of the surface of pipelines, hydro dumb, oil rigs and other strategically important objects for the purpose of early detection of corrosion, cracks, chips and other damage or changes. Also, the analysis of changes in forest tracts, geological areas and dumps in order to detect changes in the quantity and type of material being alanized.

  • 3D reconstruction for obtaining more detailed models, for obtaining a surface shape with a uniform texture, for analyzing the type of material, which greatly improves the visualization capabilities of 3D models. For 3D reconstruction, the method of shooting with a polarization filter is combined with some method of obtaining a coarse 3D model - a stereo camera or a ToF sensor.

  • Analysis of engraving on a homogeneous surface, as by conventional optical methods engraving, as a rule, is poorly discernible. For example, inscriptions on stone slabs or inscriptions on automobile tires.

  • Analysis of changes during construction, as image analysis does not allow to recognize a number of significant changes, such as, for example, wall plastering, or cracking of treated surfaces during processing.

Figure 1

Figure 2

Figure number 1 (on the right) is taken with a standard high-resolution camera. A red arrow indicates a place on which no object is visible and a smooth surface appears.
Figure number 2 is taken with the use of polarizing lens and filters and processed by our software. In this case, clearly visible is the pipe, which passes in a place marked with a red arrow.