Research On The Quality Detection For Halftone Image

The main task of this paper is to analyze and establish model for print quality detection and ink-control.In the paper, the quality detection method is introduced which bases on colorimetry, digital image processing and the most advanced color difference formula. Median Filtering is used to reduce the noise of the digital image obtained for the print. The excursion of the location for the image is regulated by the cross-line of the print. The coincidence relation between the dots on the print and the target sample is achieved by dividing a print into several areas. The correctnesss of the coincidence relation is judged by the distance in color space of two dots.After detection, the next step is the ink-control in the printing process for the rejects. In this part, ink-control is achieved by judging the dot area coverage difference between the print and target sample. The dot area coverage of print is calculated by the model established on the base of Neural Network theory which takes the L*a*b* as eigenvalues. Besides the theoretic analysis, in the process of ink-control, the Neural Network model is established and validated. We can learn by analyzing the experiment data that it is feasible to use the Neural Network model established to calculate the dot area coverage of the given area. The detection effect for the prints which printed with three kinds of chromatic ink is not as good as for the prints which printed with one kind, two kinds and four kinds of ink, but it is not a problem for the practical print process.