Method and apparatus for automatic measurement of pad geometry and inspection thereof

1. A machine vision method of inspecting of a semiconductor interconnection pad, the method comprising:

acquiring an image of the pad, the pad having a multiple zone pattern;

automatically deriving a multiple zone geometric description of the multiple zone pattern using the image by;

extracting a plurality of features from the image by extracting a plurality of line segments from the image;

generating a plurality of hypothetical geometric descriptions using the features;

selecting the multiple zone geometric description from the plurality of hypothetical geometric descriptions; and

using the multiple zone geometric description in a machine vision inspection of the pad.

2. The method of claim 1 wherein the step of extracting a plurality of line segments from the image further comprises:

extracting a plurality of edge features using a plurality of caliper tools; and

combining the edge features into clusters.

3. The method of claim 2 wherein the step of combining the edge features into clusters further comprises using a positional tolerance.

4. The method of claim 3 wherein the step of combining the edge features into clusters further comprises an acceptance threshold of a minimum number of edge features.

5. The method of claim 2 wherein the step of combining the edge features into clusters further comprises a non-maxima suppression of a group of the edge features.

6. The method of claim 2 wherein the step of combining the edge features into clusters further comprises a consistency test using a gradient direction of the edge features.

7. The method of claim 1 wherein the step of selecting the multiple zone geometric description comprises an area-based contrast evaluation.

8. The method of claim 7 wherein the area-based contrast evaluation is a normalized zone contrast.

9. A method for inspecting a multiple zone interconnection pad of a semiconductor device, the method comprising:

providing a geometric description of the multiple zone interconnection pad, the description having a plurality of zones;

providing a run-time image of a semiconductor device pad;

finding a plurality of regions in the run-time image using the zones of the description;

computing a coarse segmentation threshold in each region;

segmenting the regions using the coarse segmentation threshold corresponding to the region to provide features;

measuring characteristics of the features in each segmented region;

merging the features of each region to provide merged features; and

measuring characteristics of the merged features wherein the steps of computing a coarse segmentation threshold, segmenting, and measuring characteristics are performed for dark objects on a light background, and for light objects on a dark background.

10. The method of claim 9 wherein the step of measuring characteristics further comprises:

blob analysis of segmented features;

merging segmented features to provide combined segmented features; and

refining the combined segmented features using greyscale image data of the region.

11. The method of claim 9 wherein the plurality of regions comprise associated regions and transitional regions, and the step of segmenting the region comprises filtering the transitional regions with a median bias filter.

12. The method of claim 11 further comprising an axis-alignment of the run-time image of the semiconductor pad.

13. An apparatus for inspecting a multiple zone semiconductor pad comprising:

an image acquisition system configured to acquire an image of the multiple zone semiconductor pad, the multiple zone semiconductor pad having a plurality of regions corresponding to the multiple zones;

a machine vision system coupled to the image acquisition system, the machine vision system configured to derive a geometric description of the regions of the multiple zone pad by extracting features from the image using a caliper tool, and to combine the extracted features into clusters, generating a plurality of hypothetical geometric descriptions using the features, and selecting the multiple zone geometric description from the plurality of hypothetical geometric descriptions; and the machine vision system configured to inspect the multiple zone semiconductor pad using the geometric description of the regions of the multiple zone pad.