Research on Image Restoration and Pose Recognition Methods in Integrated Circuit Detection

Chips containing integrated circuits are semiconductor components widely used in defense and civilian sectors, with demand surging in recent years. Enhancing chip packaging efficiency is crucial for advancing the integrated circuit industry. Visual inspection is vital in chip packaging but currently limits overall efficiency, as it primarily relies on static detection methods, which are inefficient. Dynamic detection (flying vision inspection) can significantly improve efficiency, but the relative motion between the camera and chip causes image blur and trailing effects, compromising accuracy. This paper proposes a non-blind deblurring method that combines a Wiener filtering algorithm with an enhanced edge detection technique to address these issues. The method restores blurred images using Wiener filtering, followed by binarization and morphological operations. The Canny edge detection algorithm then extracts the chip’s edges, with contour analysis determining the center position. Experimental results show that the proposed method achieves high accuracy and robustness in image restoration and target localization, making it suitable for chip inspection in high-speed placement processes. This approach effectively mitigates image blur in dynamic detection, enhancing the precision and efficiency of chip placement.