机器视觉检测系统是如何在自动化包装线上进行检测的？
发布时间：
2023-02-10 11:24
产品检测是生产制造环节必不可少的一套程序，产品的合格率不仅仅是产品品质的保证，更是企业口碑宣传的必要条件。如今市场对于食品的健康安全有着高标准和严要求，人眼检测远远达不到，这时候就需要采用机器视觉检测系统来进行视觉检测。
Product testing is an indispensable procedure in the production and manufacturing process. The product qualification rate is not only the guarantee of product quality, but also the necessary condition for enterprise word-of-mouth publicity. Nowadays, the market has high standards and strict requirements for food health and safety, and human eye detection is far from reaching. At this time, machine vision detection system needs to be used for visual detection.
机器视觉检测系统在自动化包装流水线上是如何进行检测的呢?常见的食品包装流水线的视觉检测内容包括：二维码检测、生产日期保质期检测、条形码检测等，随着食品包装行业的要求越来越高，最重要的一点就是要实现产品信息可追溯。这些都需要机器视觉检测系统来完成。
How does the machine vision inspection system detect on the automatic packaging assembly line? Common visual inspection contents of food packaging assembly line include: two-dimensional code inspection, production date warranty period inspection, bar code inspection, etc. with the increasingly high requirements of the food packaging industry, the most important point is to realize the traceability of product information. All these need machine vision inspection system to complete.
目前，大部分食品生产企业都是采用喷码机或者激光打印机来印刷生产日期等信息，虽然喷码机的灵活性会比较高，但却容易出现喷印不良，比如歪了、模糊了、喷印不全的缺陷;激光打印的应用相对没有那么广泛，但却很容易出现打印重叠、漏打等缺陷，针对这些问题，如果用人工检测，就会有大的工作量，而且检测速度无法提高。
At present, most food production enterprises use ink-jet printer or laser printer to print production date and other information. Although the flexibility of the ink-jet printer is relatively high, it is prone to poor printing, such as skew, blurred and incomplete printing; The application of laser printing is relatively not so extensive, but it is easy to have defects such as printing overlap and missed printing. For these problems, if manual detection is used, there will be a large workload and the detection speed cannot be improved.
一般机器视觉系统印刷检测成像质量有很高的要求，若要使视觉系统能够正确地进行判断与决策，必须做到以下5点要求：
Generally, the printing inspection and imaging quality of machine vision system has high requirements. To enable the vision system to correctly judge and make decisions, the following five requirements must be met:
1. 摄像机要有足够的分辨率，即印品包装上的微小细节都能够清晰地再现;
1. The camera should have sufficient resolution, that is, the tiny details on the printed product package can be clearly reproduced;
2. 首先原始图像信息的质量要高;
2. First, the quality of the original image information should be high;
3. 摄像机要有足够的动态范围，即印品上的很亮和很暗的部位都能准确地反映;
3. The camera shall have sufficient dynamic range, that is, the bright and dark parts on the printed matter can be accurately reflected;
4. 摄像机的读出速度要与印品的前进速度相一致，以保证足够的精度和100%的采集。
4. The reading speed of the camera shall be consistent with the forward speed of the printed matter to ensure sufficient accuracy and 100% acquisition.
5. 摄像机要能够真实地再现印品的色彩;
5. The camera shall be able to truly reproduce the color of the printed matter;
关于机器视觉缺陷方面的检测首先采集图像，采集的方式有离线和在线两种方式，离线即事先拍摄好储存在文件夹中的图像，而在线采集就是连接相机进行实时采集图像。然后添加定位工具，对需要定位的模型进行训练，训练瓶盖作为定位模型，进行图案定位。之后再添加对比度工具，选择检测区域，设置上下限区间，对比度在区间内则OK，否则NG。
For the detection of machine vision defects, images are collected first. There are two methods of collection: offline and online. Offline means that the images stored in the folder are taken in advance. Online collection means that the camera is connected to collect images in real time. Then add a positioning tool to train the model that needs to be positioned, and train the bottle cap as a positioning model for pattern positioning. Then add a contrast tool, select the detection area, and set the upper and lower limits. If the contrast is within the range, it is OK, otherwise it is ng.
不同于人工检测员的是，我们可以将数值运用到每个缺陷，按尺寸、形状、位置、频率和发生率来量化这些缺陷，创建趋势分析数据，让生产部门能够对流程进行微调，从而改进产品质量，提高生产效率。
Unlike manual inspectors, we can apply numerical values to each defect, quantify these defects according to size, shape, position, frequency and incidence, create trend analysis data, and enable the production department to fine tune the process, so as to improve product quality and production efficiency.
随着食品行业的质量标准越来越高的法规要求下，不仅缺陷、字符、喷码需要严格的检测，条形码、二维码的阅读、验证等在许多检测过程中也变得愈发重要。
With the increasingly high quality standards of the food industry and the requirements of laws and regulations, not only the defects, characters and spray codes need to be strictly tested, but also the reading and verification of bar codes and two-dimensional codes have become increasingly important in many testing processes.
山东红宝自动化有限公司根据多年的行业经验已经服务于众多知名企业，通过对对机器视觉检测系统的现场应用，在实时监控产品生产质量、提高生产效率的同时，还可以大幅节约人工成本，是制造企业的最佳选择，欢迎致电咨询。
Based on years of industry experience, Shandong Hongbao Automation Co., Ltd. has served many well-known enterprises. Through the field application of machine vision detection system, it can not only monitor the production quality of products in real time and improve the production efficiency, but also greatly save labor costs. It is the best choice for manufacturing enterprises. Welcome to call for consultation.
机器视觉检测