An automatic system based on machine vision technology was developed for detecting surface defects of particleboard. The system was mainly composed of an industrial camera which was used for acquiring panel images and a defect detection algorithm which was implemented by C++, OpenCV,JAI SDK class library and MFC multithreading technology. The system communicated with production line control systems through computer serial ports. Test results showed that the system operated stably with a high accuracy in locating surface defects and was flexible in parameters configurations. Both of defect detection effects and operation speed of the system could meet requirements of on-line inspection for continuous press production lines.