Surface Mount Technology is the most common industrial method for manufacturing electronic devices, which involves mounting electronic components directly onto the surface of a Printed Circuit Board. Installation defects or faults that occur during the manufacturing process of these devices can potentially cause harm in certain situations. Therefore, the quality of the PCB assembly must be tested before the electronic devices are put into use. This paper proposes an automatic integrated change detection method based on unsupervised learning. This method not only has the capability to learn but also does not rely on large datasets, making it suitable for detecting installation defects of electronic components on PCBs. Experiments show that the defect detection accuracy is about 0.86, and the recall rate is 0.93. Compared to other detection methods, the execution speed is also improved, meeting the goals of practical industrial applications and achieving real application value in the industry.