| With the further promotion and deepening of the Industry 4.0 strategy and the Internet of Things technology,the QR code as an emerging Internet of Things label has gradually been applied to automated factories.The dynamic production process in the industrial environment makes the QR code tag often in the mobile environment.The relative motion between the image capture device and the target causes the QR code image to degrade,and the QR code recognition difficulty increases greatly.The high efficiency and automation requirements of industrial production oblige the QR code recognition process to be efficient and accurate.Under the stimulation of various factors,research on the QR code recognition technology in the mobile environment has become the key to promote the application of QR code.The high-precision camera can alleviate the image degradation problem to a certain extent,but its cost is unaffordable to the average small-scale enterprise.With the development of CMOS technology and the popularity of mobile smart terminals,image-reading QR code recognition methods have gradually become mainstream.The existing studies mostly focus on the identification of clear QR code images obtained at rest or at low speed,and there is less concern about the motion blur of images generated at higher speeds.For this reason,this paper studies from the two aspects of algorithm design and system development.First,a motion blur QR code recognition algorithm framework is constructed,and then a QR code recognition system under the Android platform is developed in order to achieve efficient identification of QR code.The main research work of the dissertation is as follows:(1)A rough location scheme for fuzzy QR code was proposed.The characteristics of image degradation process and motion blur QR code in mobile environment were analyzed,and a rough location scheme of fuzzy QR code based on threshold segmentation and mathematical morphology was designed.Combined with the preprocessing,the threshold segmentation,the morphological operations,the connected domain labeling,and the QR code region screening,the validity of the proposed scheme was verified by sample testing.(2)A motion blur QR code restoration algorithm was proposed.According to the general flow of motion blur image restoration,a uniform linear motion image degradation model was constructed,the parameter expression of the point spread function was determined,and the fuzzy parameters were estimated using the inverse frequency domain feature.Combined with the isotropic full variational image restoration model,the concept of image ambiguity was introduced,and an improved FTVd image restoration algorithm was proposed.The superiority of the proposed algorithm is verified by simulation comparison experiments.(3)A QR code recognition system for the Android platform was designed.The flow of QR code recognition system was designed based on the characteristics of the algorithm framework and Android platform,and the overall framework of the system was built with the concept of layered design.The QR code recognition system was developed using the Android message passing mechanism and multi-thread programming technology optimization.(4)The simulated mobile scenario verified the effectiveness of the study.Two typical relative movements and applications in industry were analyzed.Based on the existing experimental equipment,a belt conveyor experiment platform was built,an experimental scenario was set up,and the QR code recognition results were tested at different belt speeds.The superiority of this research was verified by comparing with the ZXing library. |
