CN113077449B - Image detection method for corner defects of rectangular wafer - Google Patents

Abstract

The invention provides an image detection method for corner defects of a rectangular wafer. The four preprocessed corner images of the rectangular wafer are respectively detected. First, the corner edge coordinate sequence of the corner image is obtained, and then the corner edge coordinate sequence is obtained. Judging the sequence, and finally judging whether the wafer has corner defects according to the numerical value of the judging sequence. The invention obtains the judgment sequence through the same wafer, is based on the self-comparison of the same wafer, overcomes the detection error caused by the angle difference of different wafers, and can be widely used in the detection of smooth, half-etched and fully-etched wafer corner defects Among them, it has the advantages of fast detection speed, high accuracy, and no need for training and learning, and can meet the needs of high-speed and high-precision wafer defect detection equipment.

Description

An Image Detection Method for Corner Defects of Rectangular Wafers

technical field

The invention belongs to the technical field of image detection, relates to a detection method for high-speed and high-precision defect detection equipment, and in particular relates to an image detection method for corner defects of rectangular wafers.

Background technique

In the industrial production process, it is usually necessary to use machine vision to judge whether there are defects on the surface of the product. Due to the difference in manufacturing process, batch, and degree of corrosion, the corners of rectangular wafers are not sharp, but have a certain radian and the radian varies. There are some differences. The angular arc of a rectangular wafer is not a constant value, it fluctuates within a certain range. On the other hand, the corner defects of rectangular wafers tend to be only one or two pixels in size, which is comparable to the fluctuation range of the angular arc. Corner defect detection is greatly disturbed by angular arc fluctuations. Traditional defect detection technology mainly relies on filtering and feature extraction methods, which cannot eliminate the influence of angular radian fluctuations, and it is difficult to provide effective corner defect detection results. However, the defect detection technology based on the learning algorithm has a slow calculation speed and requires a lot of training and learning, which is difficult to meet the speed requirements of high-speed defect detection equipment.

SUMMARY OF THE INVENTION

The present invention provides an image detection method for corner defects of a rectangular wafer to solve the defects of the prior art.

In order to achieve the above object, the present invention provides an image detection method for corner defects of a rectangular wafer, which respectively detects the four corner images of the rectangular wafer that have been preprocessed and judges whether the four corners of the rectangular wafer have defects. The corner vertex positions of the rectangular wafers in the preprocessed four corner images are the same, and the corner bisector passes through the midpoint of the corner image and the corner bisector is vertical or horizontal. The image detection method includes the following step:

Step 1. According to the angle bisector is vertical, choose to search the pixel points of each column of a corner image in column order, find the row number of the first pixel point that reaches the given threshold in each column, and perform all row numbers in column order. Arrange to get a sequence of corner edge coordinates; or,

According to the angle bisector is horizontal, choose to search the pixel points of each row of the corner image in row order, find the column number of the first pixel point that reaches the given threshold in each row, and arrange all the column numbers in row order to get the corner edge coordinate sequence;

Step 2: Rearrange all the data of the corner edge coordinate sequence in descending order of their serial numbers to obtain the corner edge symmetrical coordinate sequence;

Step 3: Obtain the maximum or minimum value of the data of the same serial number in the corner edge coordinate sequence and the corner edge symmetrical coordinate sequence to obtain the corner edge standard coordinate sequence;

Step 4: Calculate the absolute value of the difference between the data of the same serial number in the corner edge coordinate sequence and the corner edge standard coordinate sequence, and obtain a judgment sequence. If the maximum value of the judgment sequence exceeds a given threshold, it is judged that there is a defect;

Step 5: Repeat the above steps for the other three corner images of the rectangular wafer.

Further, in step 1, through the interpolation operation between the first pixel that exceeds the given threshold and its previous pixel, the row number of the first pixel in each column that reaches the given threshold or the first pixel in each row is obtained. The column number of the pixel that reaches the given threshold.

Further, it also includes step 6, calculating the mean value of the data of the same serial number in the standard coordinate sequence of the corner edges of the four corner images of the rectangular wafer to obtain the mean value sequence; respectively calculating the corner edge coordinate sequence and the mean value of the four corner images of the rectangular wafer The absolute value of the difference between the data of the same serial number in the sequence must be judged as the second sequence. If the maximum value of the judged sequence 2 exceeds the given threshold, it is judged that there is a defect.

The present invention obtains the result of whether the wafer has corner defects by acquiring the corner edge data of the four corner images of the rectangular wafer and performing self-comparison and mutual comparison.

The present invention has the following beneficial effects:

Even if there are differences in manufacturing processes, batches, and degrees of corrosion, the two sides of the corners of a rectangular wafer are similar, and the four corners of the same rectangular wafer are similar. The invention makes use of this feature, and obtains the result of whether the wafer has corner defects by comparing the four corner images of the rectangular wafer with each other; the invention is based on the self-comparison of the same rectangular wafer, and can adapt to different Corner differences of rectangular wafers and eliminate their errors.

In addition, the invention has the advantages of simple method, less calculation amount, and can meet the speed requirements of high-speed defect detection equipment.

The present invention will be further described in detail below in conjunction with the accompanying drawings.

Description of drawings

1 is a flowchart of an image detection method for corner defects of a rectangular wafer according to Embodiment 1 of the present invention;

Fig. 2 is the rectangular wafer in the embodiment 1 of the present invention and its four corner images that have been preprocessed;

Fig. 3 is the four corner edge coordinate series obtained in the embodiment of the present invention 1;

Fig. 4 is the symmetrical coordinate sequence of four corner edges obtained in the embodiment of the present invention 1;

Fig. 5 is the standard coordinate sequence of four corner edges obtained in the embodiment of the present invention 1;

Fig. 6 is four judging sequences obtained in the embodiment of the present invention 1;

Fig. 7 is the mean value sequence obtained in the embodiment of the present invention 1;

FIG. 8 is the second of the four judging sequences obtained in Example 1 of the present invention.

Detailed ways

In order to be able to understand the above objects, features and advantages of the present invention more clearly, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present application and the features in the embodiments may be combined with each other under the condition of no conflict.

Many specific details are set forth in the following description to facilitate a full understanding of the present invention. However, the present invention can also be implemented in other ways different from those described herein. Therefore, the protection scope of the present invention is not limited to the specific details disclosed below. Example limitations.

Example 1

An image detection method for corner defects of a rectangular wafer, the four corner images P ^k (k=1, 2, 3, 4) of the rectangular wafer that have been preprocessed as shown in FIG. Whether there are defects in the four corners, the four corner images P ^k have been preprocessed as follows: the corner vertex positions of its rectangular wafers are the same and the corner bisector is through the midpoint of the corner image and the corner bisector is vertical As an example, the steps to introduce the image detection method are as follows:

S1. Let the pixel of P ^k be

where r and c are the row and column numbers, respectively,

即

满足条件：

Let k=1, take the angle image P ^k , search from top to bottom for the row number of the first pixel that reaches the given threshold T=30 in each column

which is

To meet the conditions:

in

与其前一个像素点

的插补运算求得

Then by comparing the first pixel that exceeds a given threshold T

with its previous pixel

The interpolation operation of

按列顺序排列，得到角边缘坐标数列

all

Arrange in column order to get a sequence of corner edge coordinates

S2. Re-arrange all the data of the corner edge coordinate sequence according to their serial numbers in descending order to obtain the corner edge symmetrical coordinate sequence

S3. Take the minimum value of the data of the same serial number in the corner edge coordinate sequence L ^k and the corner edge symmetrical coordinate sequence D ^k to obtain the corner edge standard coordinate sequence

S4. Calculate the absolute value of the difference between the data of the same serial number in the corner edge coordinate sequence L ^k and the corner edge standard coordinate sequence S ^k to obtain a judgment sequence

则判断存在缺陷；If the maximum value of the judging sequence exceeds a given threshold

is judged to be defective;

S5. Set k=2, 3, and 4 respectively, execute S1 to S4 in a loop, and detect the other three corner images P ^k .

S6. Calculate the mean value of the data with the same serial number in the standard coordinate sequence S ^k of the corner edges of the four corner images, and obtain the mean value sequence A={a _c },

Calculate the absolute value of the difference between the corner edge coordinate sequence L ^k of the four corner images of the rectangular wafer and the data of the same serial number in the mean value sequence A, to obtain the second judgment sequence:

则判断存在缺陷。If the maximum value of the judgment sequence two exceeds the given threshold

It is judged that there is a defect.

Although the present invention is disclosed above with preferred embodiments, it is not intended to limit the scope of implementation of the present invention. Any person of ordinary skill in the art can make some improvements without departing from the scope of the present invention, that is, all equivalent improvements made according to the present invention should be covered by the scope of the present invention.

Claims (3)

1. An image detection method for corner defects of a rectangular wafer, which is used to detect the four corner images of the rectangular wafer that have been preprocessed respectively and determine whether the four corners of the rectangular wafer have defects, wherein the preprocessing of the rectangular wafer is performed. The corner vertex positions of the rectangular wafers in the processed four corner images are the same, and the corner bisector passes through the midpoint of the corner image and the corner bisector is vertical or horizontal. The image detection method includes the following steps: Step 1. According to the angle bisector is vertical, choose to search the pixel points of each column of a corner image in column order, find the row number of the first pixel point that reaches the given threshold in each column, and perform all row numbers in column order. Arrange to get a sequence of corner edge coordinates; or, According to the angle bisector is horizontal, choose to search the pixel points of each row of the corner image in row order, find the column number of the first pixel point that reaches the given threshold in each row, and arrange all the column numbers in row order to get the corner edge coordinate sequence; Step 2: Rearrange all the data of the corner edge coordinate sequence in descending order of their serial numbers to obtain the corner edge symmetrical coordinate sequence; Step 3: Obtain the maximum or minimum value of the data of the same serial number in the corner edge coordinate sequence and the corner edge symmetrical coordinate sequence to obtain the corner edge standard coordinate sequence; Step 4: Calculate the absolute value of the difference between the data of the same serial number in the corner edge coordinate sequence and the corner edge standard coordinate sequence, and obtain a judgment sequence. If the maximum value of the judgment sequence exceeds a given threshold, it is judged that there is a defect; Step 5: Repeat the above steps for the other three corner images of the rectangular wafer. 2. The image detection method for corner defects of rectangular wafers as claimed in claim 1, wherein in step 1, through the interpolation operation of the first pixel point exceeding a given threshold and its previous pixel point, each pixel point is obtained. The row number of the first pixel reaching the given threshold in the column or the column number of the first pixel reaching the given threshold in each row. 3. The image detection method of rectangular wafer corner defect as claimed in claim 1, is characterized in that, also comprises: Step 6, calculating the mean value of the data of the same serial number in the standard coordinate sequence of the corner edges of the four corner images of the rectangular wafer to obtain the mean value sequence; Calculate the absolute value of the difference between the corner edge coordinate sequence of the four corner images of the rectangular wafer and the data of the same serial number in the mean value sequence, and obtain the second judgment sequence. If the maximum value of the second judgment sequence exceeds the given threshold, it is judged that there is a defect.

Images