Segmentation and Features Identification for MEMS Assembly Images under Microscope

The assembly of the micro planetary gears under microscope, 400μm in diameter, is introduced briefly. To segment the outline of the target gear from the terribly degraded source image, images captured under microscope are projected into lines with different slope angles. However, the profile lacks details and important features for image understanding and object recognition but extremely decreasing the computing complexity. HSV (Hue Saturation Value) model and R-B model are applied to the analysis of the details and important features of the target. Then, based on the result of the HSV model and R-B model, target image is divided into 12 parts, and one of the gear profile is searched according to its diameter passing center point, which will offer location to the micro manipulating robot.