In an industrial context, what do we use Machine Vision for?
- Visual Inspection
- Counting
- Measurement
- Locating
- Decoding
What are some of the key factors to consider when selecting a lighting system?
- Target characteristics
- Inspection task
- Colour (or grey tone)
- Feeding speed
- Flash of continuous illumination
- Ambient conditions
What are the two main catagories of light orientations we use for machine vision?
Backlighting and Top Lighting
What are the features of Backlighting?
- Great Contrast
- Contour Projection Independent of the Object Surface
- Suitable for Dimension Measurement
What are the features of Top Lighting?
- Material Properties
- Surface Effect Image
- Sensitive to External Light
What are the five main catagories of Top Lighting used?
- Full / Partial Bright Field
- Dark Field
- Diffuse Dome
- Axial Diffuse
- Flat Diffuse
If each primary colour intensity is discretized with a byte (8 bits), then how many colours can an RGB camera detect?
(2^8)^3 = 16777216
What are the two types of Lenses?
Telecentric and Endocentric (Conventional)
What are the main features of a Telecentric Lens?
- Object size independent of distance
- No distortion, no dependent on external light
- Big lens, very expensive
What are the main features of a Endocentric Lens?
- Cheap and small
- Object size depends on the distance
- Distortions
What is the formula of the working distance of a lens?
g = f (G/B + 1)
f is the focal length
g is the working distance
G is the field view
B is the sensor size
What is the formula for the resolution?
Resolution = Length of part / Number of Pixels
What is the typical accuracy of a lens?
+5 resolution or -5 resolution
What is a Monadic Operation for Image Processing?
An operation where the output image is the same size as the input image, and the same function is applied to all pixels
What is a Diadic Operation for Image Processing?
Two images are combined into a single image, and we can use binary arithmetic operations
What is a Spatial Operation for Image Processing and what do we use them for?
Each pixel is a function of their surrounding ones
Uses include convolution, smoothing, boundary effects, edge detection, template matching etc…
What is Mathematical Morphology for Image Processing and what do we use it for?
Each pixel is a function of a subset of pixels (not only surrounding ones)
Uses include Noise Removal and Boundary Detection
