Fault inspection method

A fault inspection apparatus comprising an image pickup means for picking up an image of a specimen, a storage means for storing a reference image, and an image processing means for processing the inspection object image of the specimen picked up from the specimen by the image pickup means and the reference image stored in the storage means thereby to detect a fault on the specimen, wherein the image processing means includes:

• a displacement detection unit for detecting the displacement between the inspection object image picked up by the image pickup means from the specimen and the reference image stored in the storage means;
• a feature amount calculation unit for correcting the displacement between the inspection object image and the reference image based on the displacement information detected by the displacement detection unit and determining the feature amount of each pixel of the inspection object image and the reference image;
• a feature amount combining unit for combining the feature amounts of the inspection object image and the reference image based on the variance of the pixels of the normal portion of the inspection object image and the reference image corresponding to the feature amount of each pixel determined by the feature amount calculation unit;
• and
• a fault extraction unit for extracting a fault by determining the difference image between the inspection object image and the reference image with the feature amounts thereof combined by the feature amount combining unit.

A pattern inspection method for inspecting a pattern fault, comprising the steps of:

• acquiring a reference image by radiating light on a first one of a plurality of the same patterns formed on a specimen and picking up an image of the first pattern;
• acquiring an inspection object image by radiating light on a second one of the plurality of the patterns and picking up an image of the second pattern;
• processing the inspection object image and the reference image acquired and calculating the feature amount of each pixel of the inspection object image;
• and
• comparing the feature amount of each pixel calculated with the feature amount of other pixels and extracting a pixel having a unique feature amount as a fault candidate.

A pattern inspection method for inspecting a pattern fault, comprising the steps of:

• acquiring an inspection object image and a reference image by picking up an image of an area corresponding to a plurality of the same patterns formed on a specimen;
• calculating the feature amount of each pixel of the inspection object image by processing the inspection object image and the reference image acquired;
• mapping each pixel to the feature space in accordance with the feature amount calculated;
• setting a threshold value for extracting the pixels of fault candidates from the distribution of each pixel mapped to the feature space;
• and
• extracting a fault candidate on the feature space using the set threshold value.

A pattern inspection apparatus for inspecting a pattern fault, comprising:

• an image acquisition means for radiating light on the desired one of a plurality of the otherwise same patterns formed on a specimen and acquiring by picking the image of the desired pattern;
• a feature amount calculation means for calculating the feature amount of each pixel of the inspection object image by processing the reference image picked up from the first one of the plurality of the patterns acquired by the image acquisition means and the inspection object image picked up from the second pattern;
• and
• a fault candidate extraction means for comparing the feature amount of each pixel with the feature amount of other pixels calculated by the feature amount calculation means and extracting a pixel having a unique feature amount as a fault candidate.

A pattern inspection method for inspecting a pattern fault, comprising the steps of:

• acquiring, by a plurality of detection optical systems, images of the areas corresponding to a plurality of the same patterns formed on a specimen;
• calculating the feature amount of each pixel of the inspection object image by processing a plurality of inspection object images and reference images acquired;
• forming the feature space by mapping each pixel to the feature space in accordance with the plurality of the feature amounts calculated from the images obtained by the plurality of the detection optical systems;
• setting a threshold value for extracting the pixel constituting a fault candidate on the feature space;
• and
• extracting a fault candidate on the feature space using the set threshold value.