Modified Fusion Enhancement Algorithm Based on Neighborhood Mean Color Variation Map for AOI Solder Joint Detection

With the continuous updating of circuit board components, the demand for inspection of component welding quality is also increasing, covering the entire life process of circuit board production. In the process of detecting solder joints using AOI (Automated Optical Inspection), image processing methods are often used in combination with features to classify solder joints. In this paper, an image fusion enhancement algorithm based on the color variation matrix in the neighborhood is proposed, which extracts the dark part of the image and uses a multi-scale pyramid to fuse the images with high contrast, high exposure and color enhancement. Through this algorithm for AOI solder joint detection application scenarios, the quality of image has been greatly improved, laying the foundation for the subsequent detection process.