Halcon-Based Chip Character Recognition

The chip is used as the basis of the electronics manufacturing industry.Its surface characters contain important information such as model number of the chip and the manufacturer.In the chip production progress,it is necessary to be classified according to characters to facilitate subsequent circuit board assembly and other work.Due to the disadvantages of low efficiency and high production costs,traditional manual sorting makes machine vision an inevitable trend instead of artificial.In order to meet the higher requirements for chip sorting in the modern electronics manufacturing industry,this paper studies the algorithms related to chip character recognition and uses the machine vision software Halcon to design and implement an automatic chip character recognition system.The main tasks are as follows:The overall scheme and basic flow of character recognition are designed,including three key steps of image preprocessing,character segmentation and character recognition.Firstly,the industrial camera is used to collect the image of the chip,and the shape-based template matching method is used to locate the chip.Then the positioned chip is grayed and the image is inverted.For the case of edge diffusion in the conventional image enhancement method,the image enhancement is performed using shock filtering.The experimental results show that after preprocessing,the character area of the image is enhanced and the character edge is prominent,which lays the foundation for subsequent segmentation.For the situation that the illumination unevenness leads to different brightness in each area of the chip image,the variable threshold segmentation method is used to segment the image.The results show that the method can segment the foreground and background well even if the gray scale of the chip image changes greatly.For a single chip has fewer pixels in an image,character adhesion is caused by the edges of the characters are not clear enough,a character segmentation method based on prior knowledge is used.Experimental results show that this method can effectively segment the glued and broken characters.The commonly used character recognition methods are studied and the support vector machine is selected as the final identification scheme.The effect of different feature combinations on the recognition rate is compared.The 16-grid feature,binary feature,aspect ratio feature,foreground feature and hole feature of character are selected as the feature input of the classifier.The support vector machine classification model is constructed by using the related operator in Halcon.By comparing the combination of different parameters,the kernel function parameters with the highest recognition rate are determined and the training errors are set.Assigning 34 samples for each class of character,the classifier is trained using the samples,and chip characters are identified using the trained classifier.The experimental results show that the support vector machine has advantages in small sample training conditions,and the recognition rate can reach 92.53%.Finally,a chip character recognition system is built using Halcon and Visual Studio to realize human-computer interaction.The chip character recognition system designed in this paper has characteristics of high recognition rate,short recognition time,friendly man-machine interface,etc.It can meet the demand of electronic manufacturing for chip sorting.