Development Of Electronic Component Detection And Recognition System Based On ARM Embedded Technology

With the rapid development of the information industry,the output of electronic components is increasing continuously,and electronic components are developing toward integration,intelligence,chip actualization and miniaturization.However,this brings difficulty in component detection.Appearance detection of component is a key link in working process and its quality and pass rate directly affect the performance of electronic products using the component.Traditional visual inspection is not suitable for production line due to its false detection rate and high cost.In recent years,with the continuous improvement of embedded technology,the application combined its own advantages with image processing technology and measurement technology has been used widely.In this paper,defect detection and recognition system is developed based on ARM embedded technology combining with image processing technology.The system has the advantages of small volume,low cost,low power consumption and great real-time performance.In the system,fast and precise appearance detection is executed automatically,and then the defective component is recognized.It performs well even in narrow space and hard working consition.The main contents of this paper are as follows:(1)Building the hardware platform and software development environment of the embedded detection system.In this paper,the hardware platform is OK6410 development board based on ARM11,and the software development environment is based on Linux operating system.ARM-Linux cross-compiler environment is built on the host computer in which the Linux kernel,universal bootstrap bootloder and root file system are compiled,eventually porting Linux to OK6410 development board.(2)In this paper,USB camera is used as the input device of the detection system,image acquisition is realizd on the basis of Video4 Linux,and the image is displayed on LCD on the basis of FrameBuffer principle.(3)Certain shadow exists in the captured images which seriously affects the image segmentation and the detection and recognition of components.To solve this problem,the auto shadow detection and removal is realized by using shadow removal algorithm based on fuzzy enhancement which is proposed in this paper according to the shadow characteristic.Several digital image processing algorithms are introduced including image filtering,image enhancement,image rotation,image segmentation and edge detection etc.Specific analysis and improvement of the algorithms above-mentioned are achieved and then the image preprocessing is implemented.(4)The method for computing the smallest minimum bounding rectangle(SMBR)of detected component is designed to deal with the situation when the component pin was bended.Firstly,the angle of rotation is calculated by geometric computation,and then the component is rotated to normal position,finally,the SMBR of the component is computed.If the pin was bended,the ratio of length to width of the component changes greatly,and the ratio of length to width is certain for normal ones,so by comparing the actual ratio to the noamal value,we can screen out the defected components.Statistical pixel counting method is brought out to deal with the condition when the component pin missed.In this method,through Laplacian edge detection and binarization for the images,we count the number of the pixels of different electronic components,thus determine the threshold of pin missing component.By comparing the number of pixels to the threshold value,the pin missing component could be screened out.