Research On Image Recognition Technology Based On FDM 3D Printing

3D printing,also known as additive manufacturing technology,is a rapid prototyping technology that integrates precision machining,machinery,control and computer technology.Among all kinds of 3D printing technologies,FDM(fused deposition molding)technology is characterized by simplicity,convenience and low cost,but there are also problems such as too slow printing rate,peculiar smell generated in the molding process,and easy stacking and blocking of raw materials.In order to improve the shortcomings of FDM technology,this paper creatively adds an image recognition system on the basis of FDM 3D printing by combining the image recognition technology that has attracted great attention in recent years,so as to improve the FDM printing technology.This paper based on the independent intellectual property rights of the desktop FDM 3 d printers,first of all,its control board Marlin firmware development language,and to meet the required for image processing,the function of the main interface converts the USB to serial communication with PC connection,then studied the Cura Engine embedded in the Repetier-Host print control software by slicing tools principle,combined with the actual printing process.Then considering the experimental needs,the image preprocessing technology is studied,including several gray image transformation methods,gray image histogram enhancement,image denoising and sharpening,and united with the real-time printing image acquired by the camera for verification.After image preprocessing,it mainly uses the image edge detection algorithm to detect the real-time image in the printing process,including the first-order Sobel operator,Roberts operator,Prewit operator,and two-order Canny operator and LOG operator are studied.The principle of Canny edge detection operator and its advantages in edge detection are studied in this paper,and the image in the actual printing process is verified.Then the correlation coefficient of the image is used to analyze the printing process to distinguish whether there is a printing error.In this paper,the inherent defects in the process of FDM 3D printing are further analyzedand improved methods are proposed.An error discrimination method for a single printed object is proposed.The error detection in the printing process is realized through image detection and upper computer software control.Explored and studied the methods of path planning of FDM 3D printing,print and two objects in the process of error correcting method,using the Hilbert Peano curve for printing path of quadratic programming,and the problems in the actual printing is analyzed to solve,in combination with the feasibility and effectiveness of error correction scheme is verified by the experiments.