A Novel Quality Detection Approach for Non-mark Printing Image
In printing business, a lot of printing products have no apparent marks for registration, which cause the difficulty of printing image quality auto-detection. Aiming to this problem, a novel quality detection approach for non-mark printing image is propos
- PDF / 1,271,036 Bytes
- 8 Pages / 439.37 x 666.142 pts Page_size
- 46 Downloads / 190 Views
Shantou University Medical College, Shantou, Guangdong, China 2 Shantou Institute of Ultrasonic Instruments Co., Ltd., Shantou, Guangdong, China [email protected] 3 Shantou Polytechnic, Shantou, Guangdong, China 4 Department of Electronic Engineering, Shantou University, Shantou, Guangdong, China
Abstract. In printing business, a lot of printing products have no apparent marks for registration, which cause the difficulty of printing image quality autodetection. Aiming to this problem, a novel quality detection approach for non-mark printing image is proposed in this paper. The proposed approach mainly consists of the region feature based registration region selection and fast shape-based image matching method and an improved difference matching method to detect the printing defects. The proposed approach is realized by the well-known machine vision software HALCON. The experiment results show that the proposed approach can detect the printing defects efficiently with high accuracy, fast speed and strong robustness. Keywords: Printing image Defects detection Non-mark printing image HALCON
Registration region
1 Introduction In printing business, the quality of the products is evaluated by the degree of recovering the original pattern, the higher the better. However, in fact, for the imperfection of the printing technique and unavoidable factors, defects often appear on the surface of printing products. The traditional quality inspection is realized by manual work, which costs a great amount of time, manpower and resource. In recent years, the machine vision technique has found application in the printing inspection. The emergence of this new technique effectively solves the problem of time-consuming and poor work efficiency of the conventional manual way [1–3]. In the process of the quality detection based on machine vision technique, image registration is a step of great importance directly deciding the detecting speed and performance. Usually, the printing product has a particular mark, such as solid block with certain color, circular ring, square frame, color bar and so on. These marks can be utilized in the registration process. However, there are still a lot of printing products without apparent marks, such as cigarette brand [1]. Therefore, how to select appropriate registration regions is significant for non-mark printing products. In this paper, a novel © Springer Nature Singapore Pte Ltd. 2017 H. Yuan et al. (Eds.): GRMSE 2016, Part II, CCIS 699, pp. 173–180, 2017. DOI: 10.1007/978-981-10-3969-0_20
174
Q. Zhang et al.
quality detection approach for non-mark printing image is proposed, which is able to select registration region automatically, match speedily and detect defects efficiently.
2 The Proposed Approach The detection procedure involved two stages, detection preparation and detection, as shown in Fig. 1. Firstly, the reference image is acquired and used to generate the detection template. Secondly, the registration region based on the region features is selected and the registration template i
Data Loading...
