The work aims to realize the high-quality and high-precision recognition and positioning of the SIM chip with the machine vision technology of the chip mounter. The original image information was collected by the industrial camera and uploaded to the OpenCV system configured by VS for the image preprocessing and the reduction of image information, so as to facilitate the feature extraction. Zernike edge detection was done on the chip image to reduce the noise interference and improve the image effects. Then, the chip contour was extracted through the rectangle fitting, and the chip central coordinates were obtained. Finally, the area interval was obtained to determine the qualified chip through the extraction of chip contour area. The qualified chip was recognized and detected. The experimental results showed that, the accuracy to recognize the qualified chip reached 100% and the noise immunity was good. When the chip was positioned, the positional accuracy was within 0.5 pixel. In the complex situations, such method can ensure the mounting quality and accuracy of the chip mounter, and help improve the efficiency of qualified mounting.