Character Recognition Based on 2-D Histogram Image Threshold Segmentation
This paper describes the method and realization of the objective binary segmented image to obtain the goal of characteristic quantities. The target block of the binary image is formed by pixel labeling. Through using mathematical morphology image processing method to filter out binary image noise, it achieves a clear goal of extracting the boundary. With the license plate character recognition, the experiment shows that the algorithm is effective. All numerical examinations illustrate the high convergence speed and prove the validity of recognition rate.
L. J. Guo et al., "Character Recognition Based on 2-D Histogram Image Threshold Segmentation", Advanced Materials Research, Vols. 108-111, pp. 1366-1369, 2010