Defective part recognition device and defective part recognition method

A defective part recognition device comprising:

• a microscope for obtaining a magnified image of a unit area for recognizing a defective part on a surface of a multi-layer film substrate by irradiating the surface with white incident light;
• a spectral camera having an imaging surface on which the magnified image is formed, the spectral camera outputting spectrum information on the magnified image for each pixel of the imaging surface;
• and
• an information processing part for processing the spectrum information outputted from the spectral camera,
• wherein the information processing part includes a machine learning part for performing a clustering process on the spectrum information for each pixel, and a defect recognition part for recognizing the defective part from a process result of the machine learning part,
• the machine learning part sets a cluster according to a layer structure in the unit area and generates a histogram with a frequency that is a number of pixels clustered into the cluster, and
• the defect recognition part compares a frequency distribution of the generated histogram with a frequency distribution of a histogram free of defects, and recognizes a defective part depending on presence of a cluster having a difference in frequency.