A practical method to extract image signals in a low signal environment

A thresholding technique that can be used to extract image signals in a low signal environment is described. The technique is developed around commercially available machine vision systems. This method of determining the binarization threshold from a bimodal histogram does not require rigorous mathematical manipulation. For example, one does not need to fit two unimodal Gaussian distributions to the bimodal histogram. However, it uses the statistical mean of the histogram of the gray levels. The threshold value is used to binarize the image, which is then segmented, classified, or subjected for further image processing. The technique works also in cases of not-so-obvious bimodally distributed histograms, as in the case of low signal environment.<>