Machine Vision - Soliton Technologies

Machine Vision in Printing Industry
        Soliton Technologies, 14 year old, a pioneer in machine vision products and solutions based on India. In this August (2012)
 Soliton is going to launch a automation product that will reduce paper wastage while printing. Reduction is wastage will up to 50 percent compared to earlier  manual methods. A smart camera  would take pictures of the printed newspapers as they come out of the press, detect any errors in the color registration (CMYK). Cyan, Magenta and Yellow are secondary colours used in printing industry and K stands for Black colour. The detected errors are automatically corrected using correction motors present in the press. This process is instantaneous and thus paper wastage is reduced.

Jewellery
     They have done R&D work to develop a automation technology that will able to inspect, sort and defect detection in the manufacture of jewellery.

CASE STUDY in Pencil Manufacturing

     A pencil manufacturing   factory was manufacturing one million pencils per day. It employed 60 people inspect and remove defective pencils.  A good quality pencils, can have lead offset of 300 µm only. If the offset increases further pencil leads will break when sharpened. This length is nearly impossible to judge manually. When production was increased to 2 million pencils per day, the manufacturer went for a fully automated system. It means that within a second 23 pencils should be inspected.  This project was handed over to Soliton.  Machine Building and Automation division of Titan Industries was roped by Soliton to build system that will bring the pencils before the camera with minimal vibration. Each of the two ends are inspected by colour camera connected by IEEE 1394 interface.  Depending upon the type of pencils inspected either florescent or fiber optic light is used. They use LabView from National Instruments with IMAQ vision library provided by National Instruments (NI). Image processing algorithms have to find all defects (lead offset, wood chip-off, crack etc) within a time of 40ms. If a pencil is found defective  it is dropped into reject bin else it is lacquered, finished and boxed.

Smart Camera
Smart cameras do not require a PC for operation. It can operate stand alone. Soliton has launched a smart camera named Soliton Neo. It contains a buit-in camera, powerful DSP processor, LED lights and other required interfaces like Ethernet to connect to a laptop or remote PC. One has to use a custom software and able to control the camera thats it.

 More Info

•  Soliton is a industry sponsor for TIFAC CORE on Machine Vision  setup at Rajalakshmi Engineering College (REC), Chennai. This was inagurated by Dr. Chidambaram, Principal Scientific Advisor to the Government of India on 30th June 2009.
• Instrumentation and Control Engineering students can check analog speedometer webpage and they can try as their final year project.
• Look into Soliton webpage's case study, press release, in the news section. Very interesting project idia one may get

 Sources:

• Soliton Technologies, Bangalore, India, http://www.solitontech.com/
• Soliton Technologies launches smart camera,  http://www.hindu.com/2007/12/02/stories/2007120260951500.htm  
• System speeds pencil inspection and sorting, http://www.vision-systems.com/articles/print/volume-11/issue-12/technology-trends/quality-control/system-speeds-pencil-inspection-and-sorting.html 
• Soliton’s focus on machine vision technology, http://www.thehindu.com/business/companies/article3569849.ece 
• Analog Speedometer,http://www.solitontech.com/tmd/case-studies/automated-analog-speedometer-calibration.html

  

DIP and Sub 90nm Technology

Excerpts from the article "Comparing linear verses non linear filters in image processing"
Embedded Computing Design,
May 2012,  pp. 8-10

       Real-time or embedded image processing was limited by the cost per constraints created by the silicon substrate.  With the growth of sub 90 nano meter technology changed the entire scenario.
          Filtering is used to do interpolation, noise reduction and resampling functions. There are two types of filtering namely linear and nonlinear. As we know Image is a two dimensional signal. Filtering theory developed for one dimensional (1D) was extented to two dimentsional field. First there was analog 1D filters then came discrete and atlast digital 1D filters. A linear system should satisfy  superposition and shift invariance property. They should be causal and stable to be realized (For more information read any book on Signals and systems OR Digital Signal processing).
         Non-linear filters do not satisfy superposition and shift invariance. But they should be causal and stable to be realized. Median filters and rank-order filters are very good examples for non-linear filters.

Comment: Very nice layout, read from tablet, Thanks Mr. Joshua
http:\\embedded-computing.com/

Retina Display - Demystified

             Apple ipod touch is looking amazing because of Retina display. This makes images and movies pop off the screen. Text in pdf,  email and web pages is crisp at any size. This is because Retina display pixel density is very high and our eye is unable to distinguish individual pixels. The pixel size is mere 78 micrometer. Normal screen will be 72 pixels per inch but Retina display has 326 pixel per inch. This otherwise called as dots per inch (dpi) in printing technology. Pixels per  inch (PPI) and dot per inch (DPI) are analogous but not same. Apple has included LED backlighting and an ambient light sensor that automatically adjusts to the ambient brightness.

Good quality magazines will have 1200 DPI.  That is why always images in magazine will be very stunning.

Please check the live demo in the following website. That will help to realize yourself.
http://www.apple.com/ipodtouch/features/retina-display.html