Image analysis is closely related to image processing, computer graphics and other research contents, and overlaps with each other. The improved fuzzy C mean clustering algorithm with strong anti noise ability and neighborhood pixels is applied to the research of carbon fiber wood composites. The satisfactory results have been obtained through the test. After that the characteristic parameters of carbon fiber were extracted and a carbon fiber distribution parameter r was set up to distinguish the carbon fibers in the electron microscope. This study provides a convenient method for the identification of carbon fiber wood composites, and provides data support for the later construction of macro and micro models of carbon fiber wood composites.
