基于红外热成像的电路板载器件故障检测

Defect detection for circuit-board-mounted devices based on infrared imaging

现有红外热成像电路故障检测方法一般需要与同型号电路的标准红外热图像进行参考比对,使用范围受到限制。为此提出一种无需标准红外热图像的电路板载器件故障情况自动检测算法。算法采用生长—消除法分割提取红外热图像中发热区域,采用最小二乘法四次多项式拟合计算发热器件表面温度,配合环境温度和器件封装计算发热器件内部PN结温,从而判定器件故障。在Ubuntu 16.04和Open CV 3.0环境下进行了实验测试,验证了所提算法的有效性。

Most of the existing infrared-imaging-based methods for circcit fault detection u s a l ly need the correspond-ing standard reference infrared images for comparison,which restricts their practical application. A novel operatingtemperature and fault detection algorithm for circuit-board-mounted devices without standard reference infrared imagesis proposed. The region growing and eliminating technique is employed to extract heating areas. Then,the surface temperatures of heating devices are calculated by using least squares and quartic polynomial fitting methods. Accord-ing to ambient temperature and device package,the temperatures of internal PN calculated to judge the device fault Experiments are conducted under Ubuntul6. 04 and 0penCV3. 0 , and experimen-tal results verify the validity of this algorithm.