Creates a finder that will search the image for three finder patterns.
image to search
image to search
After a vertical and horizontal scan finds a potential finder pattern, this method "cross-cross-cross-checks" by scanning down diagonally through the center of the possible finder pattern to see if the same proportion is detected.
row where a finder pattern was detected
center of the section that appears to cross a finder pattern
maximum reasonable number of modules that should be observed in any reading state, based on the results of the horizontal scan
The original state count total.
true if proportions are withing expected limits
Like {@link #crossCheckVertical(int, int, int, int)}, and in fact is basically identical, except it reads horizontally instead of vertically. This is used to cross-cross check a vertical cross check and locate the real center of the alignment pattern.
After a horizontal scan finds a potential finder pattern, this method "cross-checks" by scanning down vertically through the center of the possible finder pattern to see if the same proportion is detected.
row where a finder pattern was detected
center of the section that appears to cross a finder pattern
maximum reasonable number of modules that should be observed in any reading state, based on the results of the horizontal scan
vertical center of finder pattern, or {@link Float#NaN} if not found
number of rows we could safely skip during scanning, based on the first two finder patterns that have been located. In some cases their position will allow us to infer that the third pattern must lie below a certain point farther down in the image.
This is called when a horizontal scan finds a possible alignment pattern. It will cross check with a vertical scan, and if successful, will, ah, cross-cross-check with another horizontal scan. This is needed primarily to locate the real horizontal center of the pattern in cases of extreme skew. And then we cross-cross-cross check with another diagonal scan.
If that succeeds the finder pattern location is added to a list that tracks the number of times each location has been nearly-matched as a finder pattern. Each additional find is more evidence that the location is in fact a finder pattern center
reading state module counts from horizontal scan
row where finder pattern may be found
end of possible finder pattern in row
true if in "pure barcode" mode
true if a finder pattern candidate was found this time
true iff we have found at least 3 finder patterns that have been detected at least {@link #CENTER_QUORUM} times each, and, the estimated module size of the candidates is "pretty similar"
the 3 best FinderPatterns from our list of candidates. The "best" are those that have been detected at least {@link #CENTER_QUORUM} times, and whose module size differs from the average among those patterns the least
Given a count of black/white/black/white/black pixels just seen and an end position, figures the location of the center of this run.
count of black/white/black/white/black pixels just read
true iff the proportions of the counts is close enough to the 1/1/3/1/1 ratios used by finder patterns to be considered a match
Generated using TypeDoc
This class attempts to find finder patterns in a QR Code. Finder patterns are the square markers at three corners of a QR Code.
This class is thread-safe but not reentrant. Each thread must allocate its own object.
Sean Owen