Hyperspectral imaging in Industrial Machine Vision
This technology is proposed by a machine vision company based in Pisa, Italy. Its systems are based on traditional vision technologies, 3D vision, spectroscopy, thermal vision and hyperspectral imaging. These systems offer remote inspection and detection capabilities which can be used to reduce scrap, automate processes, identify materials and certify products in the automotive, medical, glass, textile, robotics, ceramic tile and other spectral vision markets.
The Italian company based in Pisa offers machine vision solutions combining traditional vision and spectral vision in the form of hyperspectral vision to tackle applications in glass manufacturing, medical components identification, food processing, waste recycling and energy efficient building materials analysis.
Hyperspectral imaging was pioneered by the space industry to explore the chemical makeup of the universe. It makes it possible for the observer to remotely identify things that are not visible to the human eye. Hyperspectral technologies in the past were not inexpensive however today they have become more affordable and they are now economically more viable for deployment in numerous cost sensitive applications.
The Italian company has supplied machine vision systems, based upon PC, embedded and smart cameras solutions, to the market place since 1993 thus, with over 24 years of experience in the field of Machine vision it is perfectly positioned to combine its know-how of traditional machine vision and spectral vision to offer solutions to the relatively new hyperspectral imaging market. The technology herein proposed, example given, is in use in the “Remote Sensing with Cooperative Nanosatellites” program.
Furthermore the Italian company invests in R&D with the goal of developing new innovative products and also to improve the consultancy services which it offers. It is active in a wide range of research topics such as 3D imaging, robotic vision, hyperspectral imaging and machine vision.
Innovations & Advantages
Hyperspectral imaging generates much more spectral information than that available in the visible range of the spectrum. The main advantage of the technology which is being proposed is that it is possible to differentiate materials by their chemical composition thanks to the information present primarily in the SWIR portion of the electromagnetic spectrum (hyperspectral signatures). This is not possible with traditional machine vision systems which are based upon sensors limited to the visible portion of the spectrum. Hyperspectral technology makes it possible to deploy “easy to use” automated non-contacting machine vision systems in processes, otherwise impossible prior to the development of hyperspectral technology. Once the hyperspectral information has been identified for any given application the system setup is straight forward requiring only a supervisory involvement on the part of the end user. Payback periods, while application dependent, are short in the order of months rather than years.
This technology can be deployed in industrial processes in exactly the same manner as traditional machine systems are. The only difference is that with the hyperspectral technology much more spectral data is available for processing. These systems are compact and unobtrusive to the process. This technology can be applied in markets such as the verification of the correct composition of materials used in critical medical processes, the detection of stress cracks in glass during manufacture at very high temperatures, the determination of the presence of pesticides in crops, the separation of plastics for recycling, ripeness measurement and damage detection in foods and many more such applications.
Current and Potential Domains of Application
Hyperspectral imaging was pioneered by the space industry to explore the chemical makeup of the universe. It makes it possible for the observer to remotely identify things that are not visible to the human eye. This is now a mature technology which has found its place in many inspection tasks which were not previously possible in the visible part of the electromagnetic spectrum. This has opened an entirely new and exciting market for machine vision. Some examples of hyperspectral imaging application are described below:
• MEDICAL: Verification of the correct composition of materials used in critical medical processes where the cross contamination of products could lead to a life threatening situation.
• GLASS INDUSTRY: Detection of stress cracks in glass, invisible to eye, at the hot end of the process where the temperatures are in excess of 300C thus reducing handling risks to the quality inspectors in the factory and the end users.
• AGRICULTURE: determination of the presence of pesticides in crops.
• WASTE PROCESSING: separation of plastics for recycling.
• FOOD PROCESSING: Ripeness, damage detection