Quiz

Question: What is the purpose of the Hough transformation?

• Outlier test
• Stochastic Analysis
• Coordinate transformation
• Parameter determination of a model

Question: How many dimensions does the Hough space have?

• 1
• 2
• 3
• It depends on the model

Question: How many solutions are determined by the Hough transformation?

• Exactly one solution. The one with the most matches.
• Infinitely many solutions. The limit is set manually.
• No unique solutions. User must interactively determine best solution.
• One unique solution.

Question: What is the requirement for space resection?

• At least 3 full ground control points, if the interior orientation of the camera is known
• 3 image points
• 3 full ground control points, if the interior orientation of the camera is unknown
• Always 9 full ground control points

Question: What is the purpose of the space resection?

• Calculation of the relative orientation
• Creation of 3D objects
• Calculation of the exterior orientation of a single image
• Calculation of 3D object coordinates

Question: How are overdeterminations treated in the calculation of the space resection?

• Observations are eliminated until the system of equations can be solved unambiguously
• The observations are compensated by adjustment
• Overdetermined systems of equations can only be solved by complex algorithms and are therefore to be avoided
• The more redundant the observations, the worse the result.

Question: What is the maximum cross correlation coefficient?

• -1.0
• 1.0
• 0.0
• 0.9

Question: Does the cross correlation require approximations?

• Yes, approximations with the accuracy of half a patch size are needed
• It depends on the accuracy requirement
• No, because it is a global operator and searches the entire image domain
• Is irrelavant

Question: does the cross correlation achieve subpixel accuracy?

• No, it returns only integer solutions
• Yes, always
• Yes, when a paraboloid is fitted to the maximum peak
• There is no subpixel accuracy

Question: What is least squares matching?

• Triangulation
• Area-based matching algorithm
• Outlier test
• Search algorithm

Question: What does Least Squares Matching do?

• Global search for best gray scale match between search and sample image
• Elimination of all outliers over three sigma
• Shape matching to a square
• Adjusted matching between pattern image and given approximate position for search image

Question: What does least squares matching provide?

• Subpixel accuracy of the pattern image in the following image
• Analysis over the size of the original image
• A binary image
• Line length

Question: What is the requirement for image processing with morphological operators?

• A gray scale image
• A binary image
• An RGB image
• A 16bit image

Question: What is the purpose of mathematical morphology?

• Calculation of pixel gray values
• Estimating the accuracy spectrum
• Generate binary image
• Separation of objects for connectivity analysis

Question: What does the "Opening" method describe?

• Removal of pixel edges and a subsequent application of pixels
• Applying pixel edges
• Ablation of pixel edges
• Applying pixel edges and a following removal of pixels

Question: Do outliers influence the calculation?

• Yes
• Yes, but only with a low weight
• Yes, with a high weight
• No

Question: Does RANSAC require approximate values?

• Yes
• No
• It depends on the model
• The more accurate the approximate values, the better the result

Question: How is the outlier test done?

• Points that have deviation greater than three sigma are outliers
• Points that are out of the closest match to the total point cloud are outliers
• The point with the largest standard deviation is an outlier
• The point with the smallest improvement is an outlier