With the development of information industry and electronic technology, PCB(PrintedCircuit Board) manufacturing technology has been developed. The traditional optical microscope visual inspection method can not be applied to the precision inspection of PCB boardbecause of its defects. Precision detection based on image recognition is the development direction of modern measurement technology. In this method, the test object is imaged to a CCD camera chip by optical projection. The chip converts the test object into a digital video signal by sensitizing the image, and the geometric size of the object is obtained by analyzing the digital image. As the test is realized by computer, human interference is eliminated and the test accuracy is improved. Therefore, on the basis of constructing the PCB precision test system based on image recognition, the sub-pixel edge positioning algorithm is studied, and the PCB precision test software is developed. The test accuracy of the system is verified by the PCB test.
1. Construction of precision test system
The composition of PCB precision measurement system based on image recognition is shown in Figure 1. The system can realize the test calibration, automatic focus, large target measurement, test report result identification and output, test image saving and other functions. Working principle of the system: Firstly, the calibration work of the test system should be carried out, that is, the test accuracy of the system should be corrected according to the standard template; Secondly, the focal length of the optical lens and the working distance between the optical lens and the horizontal table are adjusted according to the specific requirements of the measurement target, and the focal length of the lens is adjusted according to the requirements of the test accuracy to determine the appropriate test magnification and accuracy. Finally, the horizontal workbench can be adjusted to realize the precision test of all the bits of the test target, especially for the test of large target objects.
2. Research on subpixel edge location algorithm
Under the condition of ensuring the quality of PCB image acquisition, the detection accuracy of PCB precision detection system based on image recognition mainly depends on the extraction accuracy of each measurement feature contour. A variety of subpixel edge detection algorithms developed in recent years can break through the limitation of physical resolution of CCD cameras, make the image edge positioning accuracy reach subpixel level, and greatly improve the image detection accuracy. The gray moment edge detection method has the advantages of simple calculation, no interpolation and iterative operation, and has high practical value. Its basic principle is to determine the location of the actual edge by assuming that the actual edge distribution in the actual image and the gray moment invariance of the ideal step edge model.
Let I(x, y) be the gray value of each pixel in the normalized edge field D of the actual image; (x0, y0) are gray barycenter coordinates of pixel points in the unit circle; S is the area occupied by h1 pixel in the edge neighborhood D; p1 and p2 respectively represent the proportion of pixels whose gray value is h1 and h2 in the neighborhood D.
3. Development of precision test software
After constructing the optical precision detection system of printed circuit board and studying the subpixel edge location algorithm, the precision detection software of printed circuit board is developed.
4. Experimental study of precision test
In order to verify the test accuracy of PCB precision detection based on image recognition, the test research was carried out on PCB precision detector, whose main test components. The test object is PCB inner board, including line width, line spacing, aperture and pad center deviation. During the test, the standard template is used to calibrate the test system, and then the circuit board is tested to get the test accuracy of the test system.
Before the test, the test system must be calibrated, that is, the standard template is used to correct the pixel value of the test, so as to obtain the specified test multiplier and the standard test value under the test unit.
Printed circuit board precision test software interface and line width test control. When testing line width, just drag the test control to the position to be measured. Edge information will be displayed automatically on both sides of line width.
The test rate was 4 times, the unit was μm, and the field of view was 6mm×4mm. After analysis, the actual width of the conductor is L0=286.2μm, the test value is the maximum width Lmax=286.8μm, the minimum width Lmin=285.8μm, then the error of the measuring system is:
The image recognition technology is applied to the precise measurement of printed circuit boards. The test results show that the system has a high accuracy (0.6μm), which can meet the high precision measurement of printed circuit boards.