Outline
Otsu's method, also known as the global optimum threshold processing using a clustering idea, the image is divided into "foreground" and "background" categories, and minimize within-class variance of these two types, namely, the maximum variance between classes.
FIG original image 1
FIG 2 using Otsu's method for image binarization
algorithm
\[ \sigma _{w}^{2}(t)=\omega _{0}(t)\sigma _{0}^{2}(t)+\omega _{1}(t)\sigma _{1}^{2}(t) \]
Wherein,
\ (T \) is divided into image "foreground" and "background" of the two threshold value, \ (\ 0} {Omega _ \) and \ (\ omega _ {1} \) are the two class weights, and \ (\ sigma _ {0} ^ {2} \) and \ (\ sigma _ {1} ^ {2} \) are the variances of these two classes.
FIG 3 FIG movable visualization algorithm
Halcon
binary_threshold(Image : Region : Method, LightDark : UsedThreshold)
When Method = 'max_separability'
the time, namely Otsu algorithm.
OpenCV
double cv::threshold(InputArray src, OutputArray dst, double thresh, double maxval, int type)
When type = THRESH_BINARY | THRESH_OTSU
the time, namely Otsu algorithm.