1. A data processing method for integrated circuit fabrication, the method comprising the steps of:

s1: acquiring an image detection image and a line laser detection image of the integrated circuit board from the image detection group and the line laser detection group;

s2: sequencing the image detection images according to the multi-area detection images set by the image detection group, and then carrying out image splicing and fusion to obtain the image detection images of the integrated circuit board;

s3: detecting an image according to the line laser; according to the scanning speed and the shooting speed, the images are spliced frame by frame to obtain a complete scanning image of the integrated circuit board

S4: importing the integrated circuit board image detection image obtained according to the S2 into Halcon software for processing to obtain a result;

s5: importing the scanning image of the integrated circuit board obtained according to the S3 into Halcon software for processing to obtain a result;

s6: calling by Halcon software to process the images of S2 and S3 to obtain a result;

s7: the results obtained at S4, S5, and S6 are derived for discriminant analysis.

2. A data processing method for integrated circuit fabrication according to claim 1, wherein: and the image detection group and the line laser detection group in the S1 are used for detecting the defects of the integrated circuit board.

3. A data processing method for integrated circuit fabrication according to claim 1, wherein: and the image detection group in the S1 comprises a plurality of optical imaging devices and at least one LED surface light source, and is used for shooting different areas of the integrated circuit board.

4. A data processing method for integrated circuit fabrication according to claim 1, wherein: the linear laser detection group in the S1 comprises at least one optical imaging device and at least one linear laser detection light source.

5. A data processing method for integrated circuit fabrication according to claim 1, wherein: the S2 further includes the steps of:

s21: extracting characteristic points of the image shot in each area of the integrated circuit board, and matching after extraction; extracting feature points by adopting an SIFT algorithm to ensure the matching precision of the images;

s22: registering the plurality of images according to the matching point set determined in the step S21;

s23: fusing the images after registration to obtain a single fused image;

s24: comparing the single fused image with a preset single calibrated standard circuit board image, and calibrating the fused image, wherein the tolerance of the circuit board appearance, holes and other mechanical characteristics is included;

s25: and finally, detecting the flaws of the integrated circuit board according to the calibrated fusion image and a learning algorithm.

6. A data processing method for integrated circuit fabrication according to claim 1, wherein: the S3 further includes the steps of:

s31: extracting characteristic points of images shot by each scanning area of the integrated circuit board, and matching after extraction; extracting feature points by adopting an SURF algorithm to ensure the matching precision of the images;

s32: registering the plurality of images according to the matching point set determined in the step S31;

s33: fusing the images after registration to obtain a single laser scanning fused image;

s34: comparing the single laser scanning fused image according to a preset single calibrated standard circuit board image, and calibrating the fused image, wherein the calibration comprises the overall dimension of the circuit board;

s35: and finally, detecting the flaws of the integrated circuit board according to the calibrated laser scanning fusion image and a learning algorithm.

7. A data processing method for integrated circuit fabrication according to claim 1, wherein: the S4 further includes the steps of:

s41: preprocessing the image of S2 to set a data set to be detected;

s42: classifying and marking the images to establish an input data set;

s43: dividing a data set, and then training and learning to construct a new classification network;

s44: and judging the defects of the image.

8. A data processing method for integrated circuit fabrication according to claim 1, wherein: the step S5 further includes the steps of:

s51: preprocessing a scanning image of the complete integrated circuit board; the preprocessing mode comprises contrast improvement and noise reduction;

s52: setting circuit boards with different defects, detecting to obtain a gray level image of the circuit board, and constructing a defect gray level image database;

s53: registering the reference gray level image database and the defect gray level image database;

s54: performing difference and threshold value screening after registration;

s55: performing matching analysis on the circuit board subjected to online detection by using the screened threshold value, and judging whether defects exist or not; if the value is smaller than the threshold value, the circuit board is qualified, otherwise, the circuit board is not qualified.

9. A data processing method for integrated circuit fabrication according to claim 1, wherein: specifically, the S6 compares the defect images of the images of S2 and S3 to obtain a comparison result.

10. A data processing method for integrated circuit fabrication according to claim 1, wherein: the S7 discriminant analysis element includes: shape, size, hole site, and circuit board defects of the integrated circuit board.

Background

Optical automatic inspection equipment, also called AOI inspection equipment, has become an important inspection tool and process quality control tool for ensuring product quality in the electronic manufacturing industry. In the SMT manufacturing process, AOI equipment automatically detects various different mounting errors and welding defects on a PCB by using a high-speed high-precision X-Y working platform and a vector characteristic visual processing technology, the minimum detectable elements are 0.1-0.05mm chip elements and IC elements with a pin pitch of 0.3mm, and the AOI equipment can be used for detecting the quality of red glue manufacturing process, tin paste manufacturing process, welding after plug-in piece wave soldering and the like. However, defects detected by conventional optical detection are also completely visible to the naked eye. The threshold setting problem is an index that needs to be considered carefully, and if the threshold is too high, erroneous judgment is likely to occur. The threshold value is too low, which easily results in missed detection. Optical inspection has great advantages for the inspection of smaller components, and is a development trend to replace visual inspection. In the manual inspection currently adopted, for a board with a large number of points or a large batch of boards, due to long-time inspection, human eyes can be fatigued, so that inspection omission is caused, and therefore, an online automatic optical inspection method for circuit board production is urgently needed to be developed, and a data processing method for integrated circuit manufacturing is urgently needed to be developed.

Disclosure of Invention

Technical problem to be solved

1) The detection efficiency of image detection and laser detection is improved;

2) the detection precision of image detection and laser detection is improved.

(II) technical scheme

A data processing method for integrated circuit fabrication, the method comprising the steps of:

s1: acquiring an image detection image and a line laser detection image of the integrated circuit board from the image detection group and the line laser detection group;

s2: sequencing the image detection images according to the multi-area detection images set by the image detection group, and then carrying out image splicing and fusion to obtain the image detection images of the integrated circuit board;

s3: detecting an image according to the line laser; according to the scanning speed and the shooting speed, the images are spliced frame by frame to obtain a complete scanning image of the integrated circuit board

S4: importing the integrated circuit board image detection image obtained according to the S2 into Halcon software for processing to obtain a result;

s5: importing the scanning image of the integrated circuit board obtained according to the S3 into Halcon software for processing to obtain a result;

s6: calling by Halcon software to process the images of S2 and S3 to obtain a result;

s7: the results obtained at S4, S5, and S6 are derived for discriminant analysis.

As a further explanation of the above solution, the image detection group and the line laser detection group in S1 are used for detecting defects of the integrated circuit board.

As a further description of the above solution, the image detection set in S1 includes a plurality of optical imaging devices and at least one LED surface light source, and is used for photographing different areas of the integrated circuit board.

As a further explanation of the above solution, the S1 line laser detection group includes at least one optical imaging device and at least one line laser detection light source.

As a further illustration of the above scheme, the S2 further includes the following steps:

s21: extracting characteristic points of the image shot in each area of the integrated circuit board, and matching after extraction; extracting feature points by adopting an SIFT algorithm to ensure the matching precision of the images;

s22: registering the plurality of images according to the matching point set determined in the step S21;

s23: fusing the images after registration to obtain a single fused image;

s24: comparing the single fused image with a preset single calibrated standard circuit board image, and calibrating the fused image, wherein the tolerance of the circuit board appearance, holes and other mechanical characteristics is included;

s25: and finally, detecting the flaws of the integrated circuit board according to the calibrated fusion image and a learning algorithm.

As a further illustration of the above scheme, the S3 further includes the following steps:

s31: extracting characteristic points of images shot by each scanning area of the integrated circuit board, and matching after extraction; extracting feature points by adopting an SURF algorithm to ensure the matching precision of the images;

s32: registering the plurality of images according to the matching point set determined in the step S31;

s33: fusing the images after registration to obtain a single laser scanning fused image;

s34: comparing the single laser scanning fused image according to a preset single calibrated standard circuit board image, and calibrating the fused image, wherein the calibration comprises the overall dimension of the circuit board;

s35: and finally, detecting the flaws of the integrated circuit board according to the calibrated laser scanning fusion image and a learning algorithm.

As a further illustration of the above scheme, the S4 further includes the following steps:

s41: preprocessing the image of S2 to set a data set to be detected;

s42: classifying and marking the images to establish an input data set;

s43: dividing a data set, and then training and learning to construct a new classification network;

s44: and judging the defects of the image.

As a further description of the above solution, the step S5 further includes the following steps:

s51: preprocessing a scanning image of the complete integrated circuit board; the preprocessing mode comprises contrast improvement and noise reduction;

s52: setting circuit boards with different defects, detecting to obtain a gray level image of the circuit board, and constructing a defect gray level image database;

s53: registering the reference gray level image database and the defect gray level image database;

s54: performing difference and threshold value screening after registration;

s55: performing matching analysis on the circuit board subjected to online detection by using the screened threshold value, and judging whether defects exist or not; if the value is smaller than the threshold value, the circuit board is qualified, otherwise, the circuit board is not qualified.

As a further description of the above scheme, the S6 specifically uses the defect images of the images of S2 and S3 to perform comparison to obtain the comparison result.

As a further explanation of the above solution, the S7 discriminant analysis element includes: shape, size, hole site, and circuit board defects of the integrated circuit board.

Detailed Description

A data processing method for integrated circuit fabrication, the method comprising the steps of:

s1: acquiring an image detection image and a line laser detection image of the integrated circuit board from the image detection group and the line laser detection group;

s2: sequencing the image detection images according to the multi-area detection images set by the image detection group, and then carrying out image splicing and fusion to obtain the image detection images of the integrated circuit board;

s3: detecting an image according to the line laser; according to the scanning speed and the shooting speed, the images are spliced frame by frame to obtain a complete scanning image of the integrated circuit board

S4: importing the integrated circuit board image detection image obtained according to the S2 into Halcon software for processing to obtain a result;

s5: importing the scanning image of the integrated circuit board obtained according to the S3 into Halcon software for processing to obtain a result;

s6: calling by Halcon software to process the images of S2 and S3 to obtain a result;

s7: the results obtained at S4, S5, and S6 are derived for discriminant analysis.

And the image detection group and the line laser detection group in the S1 are used for detecting the defects of the integrated circuit board.

The image detection group in the step S1 includes a plurality of optical imaging devices and at least one LED surface light source for photographing different areas of the integrated circuit board.

The S1 linear laser detection group comprises at least one optical imaging device and at least one linear laser detection light source.

Wherein the S2 further includes the steps of:

s21: extracting characteristic points of the image shot in each area of the integrated circuit board, and matching after extraction; extracting feature points by adopting an SIFT algorithm to ensure the matching precision of the images;

s22: registering the plurality of images according to the matching point set determined in the step S21;

s23: fusing the images after registration to obtain a single fused image;

s24: comparing the single fused image with a preset single calibrated standard circuit board image, and calibrating the fused image, wherein the tolerance of the circuit board appearance, holes and other mechanical characteristics is included;

s25: and finally, detecting the flaws of the integrated circuit board according to the calibrated fusion image and a learning algorithm.

Wherein the S3 further includes the steps of:

s31: extracting characteristic points of images shot by each scanning area of the integrated circuit board, and matching after extraction; extracting feature points by adopting an SURF algorithm to ensure the matching precision of the images;

s32: registering the plurality of images according to the matching point set determined in the step S31;

s33: fusing the images after registration to obtain a single laser scanning fused image;

s34: comparing the single laser scanning fused image according to a preset single calibrated standard circuit board image, and calibrating the fused image, wherein the calibration comprises the overall dimension of the circuit board;

s35: and finally, detecting the flaws of the integrated circuit board according to the calibrated laser scanning fusion image and a learning algorithm.

Wherein the S4 further includes the steps of:

s41: preprocessing the image of S2 to set a data set to be detected;

s42: classifying and marking the images to establish an input data set;

s43: dividing a data set, and then training and learning to construct a new classification network;

s44: and judging the defects of the image.

Wherein the step S5 further includes the following steps:

s51: preprocessing a scanning image of the complete integrated circuit board; the preprocessing mode comprises contrast improvement and noise reduction;

s52: setting circuit boards with different defects, detecting to obtain a gray level image of the circuit board, and constructing a defect gray level image database;

s53: registering the reference gray level image database and the defect gray level image database;

s54: performing difference and threshold value screening after registration;

s55: performing matching analysis on the circuit board subjected to online detection by using the screened threshold value, and judging whether defects exist or not; if the value is smaller than the threshold value, the circuit board is qualified, otherwise, the circuit board is not qualified.

Specifically, the S6 compares the defect images of the images of S2 and S3 to obtain a comparison result.

Wherein the S7 discriminant analysis element includes: shape, size, hole site, and circuit board defects of the integrated circuit board.

The working principle is as follows:

examples

The data processing method used by the invention is based on a data processing method integrating image detection and laser detection; the image detection group and the line laser detection group are used for detecting the defects of the integrated circuit board. The image detection group comprises a plurality of optical imaging devices and at least one LED area light source, is used for shooting different areas of the integrated circuit board and is used for constructing a large-pixel spliced image. The line laser detection group comprises at least one optical imaging device and at least one line laser detection light source.

Secondly, sequencing the image detection images according to the multi-region detection images set by the image detection group, and then carrying out image splicing and fusion to obtain the image detection images of the integrated circuit board; specifically, extracting characteristic points from images shot in each area of the integrated circuit board, and matching after extraction; extracting feature points by adopting an SIFT algorithm to ensure the matching precision of the images; registering a plurality of images according to the determined matching point set; fusing the images after registration to obtain a single fused image; comparing the single fused image with a preset single calibrated standard circuit board image, and calibrating the fused image, wherein the tolerance of the circuit board appearance, holes and other mechanical characteristics is included; and finally, detecting the flaws of the integrated circuit board according to the calibrated fusion image and a learning algorithm.

Detecting the image again according to the line laser; splicing the images frame by frame according to the scanning speed and the shooting speed to obtain a complete scanning image of the integrated circuit board; firstly, extracting characteristic points from images shot by each scanning area of the integrated circuit board, and matching after extraction; extracting feature points by adopting an SURF algorithm to ensure the matching precision of the images; secondly, registering a plurality of images according to the determined matching point set; fusing the images after registration to obtain a single laser scanning fused image; comparing the single laser scanning fused image according to a preset single calibrated standard circuit board image, and calibrating the fused image, wherein the calibration comprises the overall dimension of the circuit board; and finally, detecting the flaws of the integrated circuit board according to the calibrated laser scanning fusion image and a learning algorithm.

After the image detection and the laser detection data integration are finished, the subsequent analysis processing is started; specifically, the method comprises the following steps:

firstly, importing an obtained integrated circuit board image detection image into Halcon software for processing to obtain a result; firstly, preprocessing an image detection image of an integrated circuit board to set a data set to be detected; secondly, classifying and marking the images to establish an input data set; dividing a data set and then training and learning to construct a new classification network; and finally, judging the defects of the image.

Secondly, importing the scanning image of the integrated circuit board into Halcon software for processing to obtain a result; firstly, preprocessing a scanning image of the integrated circuit board; the preprocessing mode comprises contrast improvement and noise reduction; setting circuit boards with different defects, detecting to obtain a gray level image of the circuit board, and constructing a defect gray level image database; registering the reference gray level image database and the defect gray level image database; performing difference and threshold value screening after registration; performing matching analysis on the circuit board subjected to online detection by using the screened threshold value, and judging whether defects exist or not; if the value is smaller than the threshold value, the circuit board is qualified, otherwise, the circuit board is not qualified.

After the image detection and laser detection data are processed and analyzed, the steps of comparison and matching analysis are finally carried out; firstly, calling an integrated circuit board image detection image and an integrated circuit board scanning image by using Halcon software to process and obtain a result; and exporting the results obtained in the steps for discriminant analysis, specifically, comparing the defect images of the two images to obtain a comparison result. And finally, exporting final detection results obtained in the three steps, wherein the main factors of discriminant analysis in each step comprise: shape, size, hole site, and circuit board defects of the integrated circuit board.

The invention is characterized in that:

the data processing method is applied and established on the premise that a laser detection method is added on the basis of the traditional image detection, and two detection results are analyzed; compared with the traditional image detection, the laser detection has the advantages of strong environment adaptation capability and higher detection precision. The laser sensor detects some small parts, can realize contactless remote measurement, and has the advantages of high speed, high precision, wide range, strong light and electric interference resistance and the like; by combining optical imaging equipment with linear laser detection, a complete circuit board image can be constructed, and then the circuit board image is matched with the optical image more completely to realize accurate automatic detection; the method aims to provide an efficient and convenient processing method logic by combining the detection method with specific data processing, adopts different algorithm examples to preprocess, analyze and match different acquired images, can better integrate the data of the two parts by the data processing method constructed by the invention, more accurately determines the shape, the size, the hole site and the defects of the integrated circuit board of the circuit board, and is beneficial to improving the processing efficiency of the circuit board.

The control mode of the invention is controlled by manually starting and closing the switch, the wiring diagram of the power element and the supply of the power source belong to the common knowledge in the field, and the invention is mainly used for protecting mechanical devices, so the control mode and the wiring arrangement are not explained in detail in the invention.

The control mode of the invention is automatically controlled by the controller, the control circuit of the controller can be realized by simple programming of a person skilled in the art, the supply of the power supply also belongs to the common knowledge in the field, and the invention is mainly used for protecting mechanical devices, so the control mode and the circuit connection are not explained in detail in the invention.

While there have been shown and described what are at present considered the fundamental principles of the invention and its essential features and advantages, it will be apparent to those skilled in the art that the invention is not limited to the details of the foregoing exemplary embodiments, but is capable of other specific forms without departing from the spirit or essential characteristics thereof. The present embodiments are therefore to be considered in all respects as illustrative and not restrictive, the scope of the invention being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein.

Furthermore, it should be understood that although the present description refers to embodiments, not every embodiment may contain only a single embodiment, and such description is for clarity only, and those skilled in the art should integrate the description, and the embodiments may be combined as appropriate to form other embodiments understood by those skilled in the art.