The video processing functions mostly come from AForge.NET framework.
FUNCTION: parameters
Explanation
Threshold: Threshold_value
Converts the image to black and white. Pixels with brightness value smaller than the threshold_value will be black, others will be white. NOTE: Most measurements are looking for white items on black background; therefore, this is needed in almost all algorithms.
Invert: none
Inverts the image.
Filter Features by Size: Min_value Max_value
Passes only features with size falling between Min_value and Max_value. This is different than the acceptable size for the final result. You would use this mostly when looking for components based on pads recognition; with this, you can ensure that only images of pads will be considered.
Histogram: none
Increases contrast in the image.
Grayscale: none
Converts the image to grayscale.
Erosion: none
Assigns minimum value of surrounding pixels to each pixel of the result image. Removes noisy pixels, shrinks objects.
Edge detect: Method
Finds edges in the image. Four methods are provided:
1: Uses Sobel operator.
2: Calculates maximum difference between pixels in 4 directions around the processing pixel.
3: Calculates maximum difference of processing pixel with neighboring pixels in 8 direction.
4: Applies Canny edge detector.
Kill color: Radius R G B
Removes the specified color, with tolerance: Removed colors are those that are inside of RGB sphere with the specified center color and radius. NOTE: Alt+click the image to move the color of the clicked pixel to RGB parameters.
Keep color: Radius R G B
Keeps the specified color, with tolerance: Removed colors are those that are outside of RGB sphere with the specified center color and radius. NOTE: Alt+click the image to move the color of the clicked pixel to RGB parameters.
Meas. zoom: Zoom_factor
Zooms in to the image by Zoom_factor.
Blur: none
Blurs the image, reducing the effects of camera noise and possible imperfections in the target outline. Note: Computationally heavy.
Gaussian blur: Sigma
Another method to blur the image: gaussian blur with kernel size of 11. Allows for more heavy blur. NOTE: Computationally very heavy.
Hough circles: Diameter Intensity
Finds partial circles; result intensity correlates to match quality. NOTE: Computationally heavy, and not very useful.
Jog before measurement: none
If this is found on the function list, pops up an alert window, asking you to jog the machine at position before taking a measurement. Does not effect on visual video signal processing. This is useful to target fiducials on a very tight board, for example.