Image color segmentation for the inoculation of microbian cultures

The present paper describes a new color image segmentation algorithm which it has been applied with notable sucsess in the supervised automation of the inoculation process of the microbian cultures. The method is based on the 3D histogram thresholding to divide the image which is converted into three-dimensional L*a*b* color space in two diferenciated regions. In order to get a good linear segmentation a histogram stetching se realiza is made first in the mencioned color space. The algorithm is suitable to any color image type whith two diferent areas.