Testing image capture with high speed cameras

In industrial systems, objects may be moving along conveyor belts or carried on pick and place machines. A camera based inspection solution for such situations needs to take into accoun that the object in question may not be stationery, or at a fixed distance from the camera. Ergo, the captured image may not contain the object in focus, or not have the object in view at all. One option would be to stop the conveyor belt to take a photo of the object:

 


However, stopping a conveyor belt is not always possible. You would need cameras that can capture a clear image even when objects are moving. Enthusiasts naiively assume that one can take a video and perform analysis on individual frames of the video. Unless the video has the correct frame rate and good resolution, the solution would not work. Assume that you need to implement an industrial imaging solution to monitor items moving on a conveyor belt. You should consider specifications such as these:
  • Distance from camera to conveyor belt? Example: 3 meters
  • Speed of the conveyor belt? Does this have a range? For example: 0.8-1.4 meter per second. This turns out to be about 4-6 KMPH 
  • What kind of lighting can be placed? 
Given these parameters, one would need to test high-speed cameras. We tested a high speed camera (Basler aCA 1300-200 micrometer) by using a calibration chart. The chart was attached to a steel plate connected to a DC motor rotating at adjustable speeds. We examined the result at different RPMs and distance settings. Each video is of a few seconds duration. The camera is set to 130 fps and several images can be extracted from each video. Sample videos are below:
 



Two frames are shown below for comparison:











Analysis from our examination:
  • FoV analysis (Distance from camera to object vs area of the object observed): 
    • At 200 mm, we are able to observe 30 mm x 25 mm 
    • At 600 mm, we are able to observe 100 mm x 80 mm 
    • At 1000 mm, we are able to observe 150 mm x 120 mm 
  • When the camera is placed at 600 mm distance from the object, details less than 3 mm wide are out of focus in most frames. At 600 mm, the camera resolution is not capable of viewing details less than 3mm. 
  • At 100 mm, the camera is unable to focus on the object placed in front. 
  • At both 2 and 5 RPM, about 30% of the frames were out of focus.

Comments

Post a Comment

Popular posts from this blog

Discrete Wavelet Transform

Code and sample files for this demo can be downloaded from here . Code uses only libraries from the ACV-VirtualMachine.No extra libraries are needed. You can unzip the code to any folder on your system. All paths used in the code are relative. We are using the folder "~/acv/code/dwt_demo/" Compiling and running the code is fairly straight forward. You only need to run the following two commands: acv@acv-vm : ~/acv/code/dwt_demo/bin $ cmake ../src/wavelet_demo_others/ -- The C compiler identification is GNU 7.5.0 -- The CXX compiler identification is GNU 7.5.0 -- Check for working C compiler: /usr/bin/cc -- Check for working C compiler: /usr/bin/cc -- works -- Detecting C compiler ABI info -- Detecting C compiler ABI info - done -- Detecting C compile features -- Detecting C compile features - done -- Check for working CXX compiler: /usr/bin/c++ -- Check for working CXX compiler: /usr/bin/c++ -- works -- Detecting CXX compiler ABI info -- Detecting CXX compiler ABI info - ...
Read more

Geometric image transforms

Image
 The sample code and data for this tutorial can be downloaded from here . The result files are here . The accompanying slide deck for this tutorial is here . After downloading the code, run the following commands:  acv@acv-vm : ~/vision/practice/bin $ cmake ..; make; ./homework transformDemo -- The C compiler identification is GNU 7.5.0 -- The CXX compiler identification is GNU 7.5.0 -- Check for working C compiler: /usr/bin/cc -- Check for working C compiler: /usr/bin/cc -- works -- Detecting C compiler ABI info -- Detecting C compiler ABI info - done -- Detecting C compile features -- Detecting C compile features - done -- Check for working CXX compiler: /usr/bin/c++ -- Check for working CXX compiler: /usr/bin/c++ -- works -- Detecting CXX compiler ABI info -- Detecting CXX compiler ABI info - done -- Detecting CXX compile features -- Detecting CXX compile features - done -- Found PkgConfig: /usr/bin/pkg-config (found version "0.29.1") -- Found OpenCV: /usr/local (found...
Read more