| In the product packaging industry,carton boxes usually use the hot melt glue machine to spray hot melt glue on the two pieces of cardboard for bonding.But the hot melt glue machine is a complex linkage system.The hot melt glue gun is easily blocked by impurities causing the hot melt glue to not spray properly,causing quality problems in the carton packaging process.This may result in the product being damaged during circulation.To prevent the cost loss caused by hot melt glue spraying quality from the beginning of product distribution,it is necessary to test the quality of hot melt glue application.The traditional methods of manual sampling and temperature sensor detecting are inefficient,inaccurate,and subjective on the one hand,and require a very demanding installation environment on the other.In addition,the hot glue is sprayed between two pieces of cardboard,which is not directly detected by the machine vision system based on ordinary cameras.Therefore,it is necessary to introduce a long-wave infrared camera to capture the temperature characteristics of the hot melt adhesive and to study a multi-spectral composite imaging-based hot melt adhesive coating quality inspection system.The main contents of this paper are as follows.(1)The demand for the hot melt adhesive coating quality inspection system is analyzed.Five possible defects under actual working conditions are also analyzed.A vision inspection scheme based on multispectral composite imaging is developed,and the equipment hardware is designed and selected.(2)Aiming at the problem that visible and long-wave infrared cameras are difficult to be calibrated at the same time because they capture information under different spectra,a multispectral calibration plate and calibration method that can be applied to both visible and long-wave infrared cameras are designed to solve the following problems: 1)the existing calibration plate cannot act on both long-wave infrared and visible cameras;2)The thermal interference of the calibration image of the long-wave infrared camera is serious.(3)The corresponding image processing and analysis methods,including image denoising,image enhancement,pseudo-color processing of infrared images,ROI selection,and feature extraction,are studied for the target images acquired by the visible camera and the long-wave infrared camera.(4)The detection system proposed in this paper is implemented and analyzed,and the specific program steps for hot melt adhesive coating quality detection under actual working conditions are given.And the specific scheme steps are given for the detection of each type of hot melt adhesive coating quality defects.The experimental analysis is also carried out,and the average accuracy reaches 97.6%,which meets the needs of the actual working conditions. |
PDF Full Text Request