The stability of PCB signals is crucial in electrical control in various fields such as military, aerospace, etc. Incorrect pin positions can cause equipment damage. For complex electrical control demand scenarios, PCBs require multiple sets of power lines to supply power together. The power supply pins are distributed in specific positions on the large circuit board. During the manufacturing process, due to the miniaturization of the circuit board and the dense arrangement of pins, problems such as pin missing, misalignment, and disconnection are prone to occur. This article proposes an online detection technology for pin defects. Preprocess the color image into a grayscale image using weighted average method to improve processing speed. Use maximum entropy global threshold segmentation and multiple morphological processing to accurately obtain the middle row pin area, greatly preserving parameters such as pin length and area. Compare and distinguish with standard long and short pin parameters to achieve pin missing and disconnection detection. In the tested image samples, the system can locate the power supply pins in real time and accurately identify pin defects.