Real-Time 3d Image Fusion Based on Wavelet Transform in Microscope Auto-Focusing System

With the limitation of the depth of focus in microscope, especially in high-power objectives, different areas in the field of view have different focal planes and therefore only sharp focused images in a small part of the view can be obtained. A three-dimension micro-image fusion algorithm is introduced based on wavelet transform (WT) in microscope auto-focusing system. The step motor is driven continuously moving across the vicinity of the focal plane to obtain the micro image information of the relevant position. With real-time multi-focused images acquired and saved, the image fusion algorithm is applied to finally get a complete and sharp three-dimension image. This can be used to extend the depth of focus in microscope and in three-dimension reconstruction in biologic slice, rock specimen, and so on. The particularity brought by the image fusion of a real-time system is mainly studied as well as the algorithm based on this condition. Here we emphasize the multi-focus fusion algorithm based on WT. Images acquired from the digital camera are transformed from spatial domain to wavelet domain. All the details from different images are extracted and processed from the wavelet coefficients and then are recombined to form the new coefficient of the resulting image. By taking an inverse WT, we finally get the final three-dimension image. This system has two main advantages: real-time processing and accurate fusion result.