JPS59973B2 - Automatic visual inspection equipment - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to an automatic visual inspection device for semiconductor pellets (diode pellets) and the like.

Conventionally, defects such as pellet chips, surface abnormalities, and cracks in semiconductor devices such as transistors and diodes are generally visually inspected. In an example of an automatic visual inspection device, a method is adopted in which the area, perimeter, length of the sides of a circumscribed quadrilateral, etc. are measured to discriminate between non-defective products and defective products. Although this determination method can sufficiently detect large defects or defects of a specific shape, surface contamination inside the pellet,
Fine defects such as surface plating peeling were difficult to detect because the degree of change was small. The purpose of the present invention is to eliminate the above-mentioned drawbacks of the conventional technology, and to avoid problems such as minute surface plating peeling and surface contamination when inspecting the appearance of two-dimensional patterns with uniform and convex surfaces, such as semiconductor pellets. An object of the present invention is to provide an automatic visual inspection device capable of detecting isolated portions inside a pattern so as to sensitively detect even defects.

That is, in the present invention, if there is no defect in the pattern, in an outwardly convex figure, there are only two pattern edges in one scanning line; When a point exists, we focus on the existence of three or more pattern edges, and detect the existence of an isolated defect by detecting how many times a single scanning line encounters a pattern edge when scanning the pattern. It was decided whether or not to do so. However, when checking a single scanning line, even if an isolated point that is not originally intended for one bit occurs within or outside the pattern due to some kind of disturbance, this point is determined to be a defect. Therefore, a method is adopted in which the size of the defect is arbitrarily set and it is determined that a defect exists only when an isolated point with a size equal to or larger than the set value is found. Of course, it is also possible to set the size of the defect to a minimum of 1 bit. As a method for determining the size of an isolated point, the size in the direction parallel to the scanning can be determined by considering that when a plurality of pattern edges exist within a set distance, they are regarded as one edge.

In the direction perpendicular to scanning, when three or more pattern edges exist in a set number of consecutive scanning lines, it is determined that there is a defect. In this way, it is possible to set the size and detect isolated defects. A specific example of the present invention applied to the inspection of a graphic pattern 1 as shown in FIG. 1 will be described. In the present invention, the test pattern is described as a square, but any convex pattern may be used. 2 to 6 are scanning lines,
The video signals shown in FIGS. 7 to 11 correspond to the state of graphic pattern 1.

If a pulse is generated at the edge of the figure and counted, the counter values for scanning lines 2 to 6 will be 2, 4, 4,
It will look like 4,4. A value other than 2 indicates that an isolated point exists on the scanning line.

The size of an isolated point can be determined by how many consecutive scanning lines have a counter value other than 2.

FIG. 3 shows an example of a hardware configuration for detecting the presence of isolated points in N or more consecutive scanning lines. 12 is an imager that captures an image while scanning the pattern 1 to be inspected; 13;
Reference numeral 14 indicates a binarization circuit that binarizes the video signal obtained from the image pickup device 12, and an outline extraction circuit 14 that extracts the outline of the pattern to be inspected from the signal that has been binarized by the binarization circuit 13.

Then, the binarized video signal is differentiated to extract the contour signal,
A counter comprised of flip-flops 15 to 17 detects whether or not there are three or more pattern edges in one scanning line.

This counter is cleared by a scan start pulse 19 issued at the start point of the scan line, that is, at the left edge of the two-dimensional screen, and is ready for new counting. The value of the counter is determined to be 3 or more using the scan end signal 20 that is output when the scan line reaches the right edge of the screen at the end of the scan line. counter 2 as existing in
Add one to 1. If there is no isolated point when the scanning line reaches the right edge of the screen, the output 24 of the gate 18 becomes 01 since the number of edges is 2 or less, and the scan end signal 20 is passed through the delay circuit 25. The delayed signal 22 passes through the gate 23 and clears the counter 21. Therefore, whether the counter 21 is within the set value N or not is made so that the counter overflows when it exceeds N.
If an overflow (0verf10w) is detected, it indicates that an isolated point exists across N scanning lines. Strictly speaking, in this detection method, even if isolated points exist at intervals, if they span N scanning lines, the result is that one isolated point occupies the width of N scanning lines. The result will be similar, but this is not a problem in practical terms. As described above, according to the present invention, it is possible to determine whether or not an isolated minute defect exists in a two-dimensional figure in time series with a simple configuration without storing the video signal output from the imaging device. It has the effect of being able to be detected and also to be processed at high speed.

Furthermore, the device configuration is only a combination of two or three counters and gates, and is highly effective as a defect determiner when automating the visual inspection of two-dimensional convex shapes such as semiconductor pellets.

[Brief explanation of the drawing]

FIG. 1 is a diagram showing an example of the pattern of an outwardly convex object to be inspected, which is the object of the present invention, and FIG.
The video signal when the pattern shown in the figure is captured by an imager is 2.
FIG. 3 is a diagram showing a video signal obtained by converting into a value, and is a diagram showing a schematic configuration of an embodiment of the automatic visual inspection apparatus of the present invention.

Claims (1)

[Claims] 1. An imaging device that sequentially scans a two-dimensional figure in time series and converts it into a video signal, and 2.
A binarization circuit that converts into a digitized signal and the above 2 for each scanning line.
a first counting means for counting the number of times the binary video signal outputted from the digitization circuit is inverted; The first comparing means for making a comparative judgment and the first comparing means detecting that the value output from the first counting means is equal to or greater than a specified number over consecutive scanning lines. A second counting means for counting the number of scanning lines and a second counting means for comparing and determining whether or not the number of scanning lines counted by the second counting means is equal to or greater than a preset specified value. 1. An automatic appearance inspection apparatus comprising: a comparison means, and is configured to determine whether or not an isolated defect exists in a two-dimensional figure based on a signal obtained by the second comparison means.