Research on Vision-Based Chip Pose Measurement Method
Aiming at the problem of chip offset detection during chip mounting,a vision-based chip pose measurement method is proposed.Firstly,the chip image is preprocessed.On the basis of pixel level edge,according to the gray value of pixel point,Sigmoid function is used to classify the chip sub-pixel edge accu-rately,and the edge outliers were corrected.Finally,the number of chips in the image is obtained by count-ing the connected components in the image according to the known chip size,the number of chips in the image is determined,and the chip sub-pixel edge is divided into four areas by the center of gravity coordi-nates,and the least square method is used to fit sub-pixel edge respectively,the pose information of the chip was obtained.The experimental results show that the method has certain robustness,and the positioning ac-curacy is less than 0.5 pixel,and the running time is about 0.4 s,which meets the requirements of mount-ing in actual production and has certain application value.
NETL
NSTL
万方数据
维普
