| As the rapid development of information technology and automatic control technology, combining the two techniques and applying them to traditional handicrafts, manufacturing and processing industries has become an important research topic. In the traditional injection molding industry, the original hand laminated and stripping mold gradually developed into mechanical laminated and stripping mold. But in the process of plastic products’stripping, there will appear stripping which is not completed enough, slightly residual, or foreign matter, it requires manual intervention, which results in that robot can not fully play the role, so the focus of this study is to find an effective way to solve the problem.The purpose of this study is the use of embedded machine vision technology to rapidly detect the residues during the injection molding production. The automatic detection technology replaces manual care, achieves the plastic stripping process monitoring, increases the automation of injection, and reduces the loss of equipment. The system uses an embedded hardware platform based on TQ2440, and is combined with embedded image processing algorithms to achieve a highly automated intelligent detection system. This major research work is as follows:1. According to the detection system requirements, hardware platform selected TQ2440. Analyzing and comparing the advantages and disadvantages of several embedded operating systems, the embedded Linux system is selected as a platform operating system. The system’s overall design is based on that.2. Studying of images preprocessing method, one befitting image preprocessing techniques is selected for the system. Analyzing the deficiencies in the process of smoothing, the improved method is proposed, which makes more use of the pretreatment image for the following work running smoothly. Analysis of the environmental which impacts the system, such as light, shape of the mold or other factors, an exact image threshold is presented through trial.3. Focusing on analyzing and studying the image processing algorithms of the detection system, and combining with the actual situation of the system, the poor shadow algorithm of the image is selected as the basic algorithm. The existing poor shadow algorithm has been improved to improve detection efficiency of the system, and make the test results more accurate.4. Analysis of the system’s hardware platforms peripherals supported by appropriate equipment and image acquisition devices OV9650module, analyzing Bootloader porting, the migration of the embedded Linux system and the rebuilding of the file system, the driver of the system’s camera module is programmed.5. The system’s application software is designed and downloaded to the hardware platform through cross-compiler tool. Combining with the mechanical part, the joint commissioning of the equipment is tested. Finally, the tests come on the existing injection molding machine and the collected data of the system were analyzed.The innovation of this paper is that the embedded machine vision technology is used in the product residue testing of injection molding industry, and that detecting injection product residue is with the use of image pre-processing technology and the improved poor shadow algorithm. Considering of environmental factors to the injection site, some of the compensation use the light compensation method, which can be more effective and fast to improve detection rates. In addition, the detection system using the existing features of injection molding machine, without any major changes, can be installed directly into existing equipment to do detecting. Through experimental testing and analysis, and associating with the analysis results of collected image data show that the performance of the system basically reaches the initial target. It has a certain role in the entire injection molding industry. |
PDF Full Text Request