On-line printed matter color quality evaluation method based on probability neural network algorithm

Abstract

The present invention discloses an on-line printed matter color quality evaluation method based on a probability neural network algorithm. The method comprises the steps of selecting a printed matter color quality evaluation index, establishing a probability neural network algorithm model, utilizing a matching on-line detection device to obtain the color quality evaluation index of a to-be-measured printed matter, and inputting the obtained index data to the algorithm model, so that the probability neural network algorithm model can output the printed matter color quality evaluation result by calculating. The probability neural network modeling time of the method only needs 0.041313 seconds, the new training samples can be obtained online and real-timely, the quality evaluation index also can be modified according to the demands, and the algorithm model can be constructed rapidly. The method is high in fault tolerance, and the printed matter color quality evaluation result can be optimized by just obtaining more training samples online and real-timely and inputting to the algorithm model. The method is stable in rating, and the conformity of the evaluation result and a human eye comprehensive evaluation effect is very high.

Description

technical field [0001] The invention relates to an online print color quality evaluation method, in particular to an online print color quality evaluation method based on a probability neural network algorithm. Background technique [0002] At present, in the printing production process, the printing machine personnel need to conduct random sampling inspections on color prints, and evaluate their quality according to the corresponding quality evaluation indicators. When evaluating, it is necessary for the machine personnel to make comprehensive judgments based on their own experience, which leads to poor repeatability of the comprehensive judgment results. In order to solve this problem, a specific algorithm has been used to simulate the process of comprehensively evaluating the color quality of printed matter by human eyes. The evaluation method ultimately ensures that the color quality of printed matter can always reach the optimal standard, and the essence of comprehensiv...

AI Technical Summary

Problems solved by technology

[0002] At present, in the printing production process, the printing machine personnel need to conduct random sampling inspections on color prints, and evaluate their quality according to the corresponding quality evaluation indicators. When evaluating, it is necessary for the machine personnel to make comprehensive judgments based on their own experience, which leads to poor repeatability of the comprehensive judgment results. In order to solve this problem, a specific algorithm has been used to simulate the process of comprehensively evaluating the color quality of printed matter by human eyes. The evaluation method ultimately ensures that the color quality of printed matter can always reach the optimal standard, and the essence of comprehensive evaluation of the color quality of printed matter by human eyes is to first obtain the comprehensive evaluation indicators of each printed matter, and then combine the rating standards of each evaluation index through the human brain to perform complex The calculation, and finally output the color quality evaluation results of printed matter, this process belongs to the process of pattern recognition, and the artificial neural network is the best method for pattern recognition. Currently, the algorithms used to simulate the comprehensive evaluation of human eyes include BP artificial neural network algorithm and Fuzzy artificial neural network algorithm, but all have the disadvantages of slow convergence speed and non-optimal convergence results

Images