Automatic Leveling System for LED Die Bonder Based on Sub-pixel Edge Detection

To address the issue of LED die bonders being unable to detect and correct minor displacements in curved surface environments, resulting in low chip placement accuracy and accumulating deviations, this paper proposes an automatic leveling system based on sub-pixel edge detection. On the software side, the Zernike moment sub-pixel detection method is employed to extract circuit edges and monitor pixel width variations in real time, thereby inferring the displacement of the chip relative to the die bonding position. On the hardware side, a stepper motor drives the rotary axis, and the calculated correction angles are converted into PWM pulses, forming a closed-loop leveling control system. Experimental results demonstrate that the automatic leveling system can stabilize angular deviations within ±0.1°, achieving rapid and precise correction of chip placement. This solution offers significant advantages in reducing manual intervention and increasing yield rates in large-scale production, providing technical support for high-precision automated applications of curved surface die bonders.