A laser speckle detection image enhancement method fusing infrared images

Abstract

The application discloses a kind of laser speckle detection image enhancement methods of fusion infrared image, first, using the advantage that infrared thermal imaging technology monitors object temperature change in real time, through transmission transformation, affine transformation or single mapping matrix etc. Method is registered to the same coordinate system with speckle image and infrared image, and the position of defect is quickly positioned;Second, using the method of sequential image processing, the image data on the whole time sequence is analyzed, and more defect information can be obtained;Third, Fourier phase analysis is carried out on the sequential image, and the background noise is further removed, the phase diagram obtained by Fourier phase transformation is obtained by morphological operation, and the defect boundary is extracted from the original shear speckle sequential image data. Each defect stripe with the best defect quality is extracted, the phase enhancement of shear speckle image is realized, the contrast of speckle stripe image is significantly improved, and the purpose of image increase is achieved.

Description

Technical Field

[0001] This invention belongs to the field of optical nondestructive testing, specifically a laser speckle detection image enhancement method that integrates infrared images. Background Technology

[0002] In recent years, numerous scholars both domestically and internationally have conducted extensive research on non-destructive testing methods for composite materials. In the aerospace, shipbuilding, and military industries, most composite materials used are bonded structures. Current testing methods primarily include ultrasonic testing, X-ray testing, laser shear speckle interferometry, and thermal imaging non-destructive testing.

[0003] Infrared thermal imaging, as a non-destructive testing technology for acquiring and processing thermal signals, allows heat to propagate through the surface of an object in the form of thermal waves when thermal excitation is applied. When the thermal waves reach abnormal areas, the propagation rate changes. When there are defects such as fractures, debonding, or delamination inside the object, temperature differences will appear between different areas. It can monitor the changes in the temperature field of the object surface in real time and observe the location of defects. However, defects in the original infrared thermal image have problems such as blurred edges and low contrast.

[0004] Laser shear speckle nondestructive testing uses coherent optical interferometry to measure the partial derivative of the out-of-surface displacement of an object's surface under stress. It boasts advantages such as wavelength-level sensitivity and full-field measurement, as well as being non-contact, high-resolution, and highly accurate. It is highly effective in detecting near-surface debonding defects in materials. When an object has surface or internal defects, its phase image will exhibit a high-contrast butterfly pattern.

[0005] When laser shear speckle interferometry is used to detect debonding or rubber-rich areas in rubber-metal bonded structures, after long-pulse thermal excitation, the quality of the fringes decreases as the defects become denser with deformation, especially when the defects are large or shallow. As the fringe order increases, the fringe information at the defect becomes close to noise, making it impossible to observe butterfly patterns with good contrast, thus reducing the defect recognition rate. Furthermore, the overall phase image of the shear speckle also exhibits significant noise that changes over time, making it impossible to obtain high-quality fringe information about the defect from a single phase image at a given moment. Summary of the Invention

[0006] The purpose of this invention is to address the shortcomings of existing laser shear speckle technology and, by combining the advantages of infrared thermal imaging technology for real-time monitoring of object temperature changes, propose an image enhancement method for laser shear speckle interferometric nondestructive testing that fuses infrared images. This method can obtain high-quality fringe information of all defects in a single speckle phase image.

[0007] This invention discloses an image enhancement method for laser speckle detection by fusing infrared images, comprising the following steps:

[0008] A laser speckle detection image enhancement method that fuses infrared images includes the following steps:

[0009] (1) Place the sample on the optical platform with the detection surface of the sample perpendicular to the camera field of view. At the same time, set up the speckle interferometer and infrared equipment to make the sample fill the field of view as much as possible. Adjust the shearing mirror to a suitable shearing angle so that the two overlapping images on the imaging surface present a certain amount of misalignment. The initial shearing angle can be set to about 45°. Measure the deformation produced and register the speckle image and infrared image by methods such as transmission transformation, affine transformation or monochromatic matrix.

[0010] (2) Apply load to the sample by heat loading to cause it to debond or the glue-rich local area to produce relative deformation. Use infrared equipment to monitor the surface temperature change of the object in real time, obtain the location of the defect and extract the geometric center of the defect. At the same time, record the deformation diagram by speckle interferometer.

[0011] (3) The obtained speckle deformation map of the sample is automatically filtered and dephased to obtain the phase change value generated by the defect location of the sample, and the speckle sequence phase map is acquired.

[0012] The actual deformation caused by the defect location in the sample is obtained from the formula for the optical path difference based on the phase change value, where the relationship between the phase change value and the actual deformation value is:

[0013]

[0014] In the formula, w represents the relative deformation generated by the two detection surfaces, and λ represents the wavelength of the laser. This represents the phase change value;

[0015] (4) Perform a one-dimensional Fourier transform preprocessing operation on the phase map of the speckle sequence acquired in step (3) to obtain the phase map at the first frequency in order to further remove background noise. In order to improve computational efficiency, the rectangular region with defects in the speckle image is taken as the region of interest, and only the region of interest is subjected to a one-dimensional Fourier transform operation.

[0016] The expression for transforming an image from the time domain to the frequency domain, that is, transforming f(x) into F(ω), is:

[0017]

[0018] (5) Filter and segment the phase image obtained in step (4) to obtain the binary image of the defect and the boundary of the defect. This includes the following steps:

[0019] (5.1) Perform median filtering on the phase map at the first frequency obtained after one-dimensional Fourier transform to further filter out the speckle noise in the speckle phase map.

[0020] (5.2) The filtered phase map is segmented using a fixed threshold to obtain a binary map. Then, the binary map is opened to eliminate noise and small connected components in the image. Then, the area reduction of larger connected components caused by erosion is compensated by dilation operation. After this operation, the approximate boundary of the defect can be obtained.

[0021] (6) Taking advantage of the fact that the phase of the speckle phase map is all in (-π,π), the phase value at a certain position changes cyclically between (-π,π). When the phase value at that position changes abruptly, a new first-level fringe is generated. Track the gray value change of a pixel near the geometric center, and use the pixel difference between two adjacent moments to represent the derivative of the pixel gray value function. When the derivative changes abruptly, a new first-level fringe is generated. Use the number of times the derivative changes abruptly as a threshold. By setting an appropriate threshold, the phase map of the current defect when it is in a low-level fringe can be extracted by the position and boundary of the defect geometric center.

[0022] (7) By fusing the low-level stripe patterns of all defects into a single phase image, a clear and high-quality phase image can be obtained in the same image, thus achieving the purpose of image enhancement.

[0023] Compared with existing technologies, this invention firstly utilizes the advantage of infrared thermal imaging technology, which can monitor the temperature changes of objects in real time. By using methods such as transmission transformation, affine transformation, or monochromatic matrix, speckle images and infrared images are registered to the same coordinate system, allowing for rapid location of defects. Secondly, by employing sequential image processing methods to analyze image data over the entire time series, more defect information can be obtained. Thirdly, Fourier phase analysis is performed on the sequential images to further remove background noise. Morphological operations are then used on the phase map obtained after Fourier phase transformation to obtain defect boundaries. Finally, by combining the original sheared speckle sequence image data, the defect stripes with the best quality for each defect are extracted, achieving phase enhancement of the sheared speckle map and significantly improving the contrast of the speckle stripe image. Attached Figure Description

[0024] Figure 1 This is a flowchart of the image enhancement method of the present invention;

[0025] Figure 2 These are photographs of the rubber-metal bonding structure specimens used in the embodiments;

[0026] Figure 3 This is a schematic diagram of the laser shearing speckle interference optical path used in the embodiment;

[0027] Figure 4This is a schematic diagram of the device and its arrangement used in the embodiment;

[0028] Figure 3 and Figure 4 In the middle, 1. Sample 2. Laser 3. Beam splitter 4. First reflecting mirror 5. Second reflecting mirror 6. Phase shifter 7. Camera 8. Heating lamp 9. Laser shearing speckle interferometer;

[0029] Figure 5 This is a phase map of the speckle sequence defect state changes acquired by software after deformation was applied in the embodiment.

[0030] Figure 6 This is the frequency domain phase map obtained by performing a one-dimensional Fourier transform on the sequence phase map in the embodiment.

[0031] Figure 7 This is a binary image of speckle defects obtained after a series of morphological operations in the embodiment.

[0032] Figure 8 This is a fused image obtained by extracting and fusing defects in low-level stripes into a single image. Detailed Implementation

[0033] The present invention will be further described in detail below with reference to the embodiments and accompanying drawings, but this is not intended to limit the present invention.

[0034] Example

[0035] The rubber-metal bonding structure sample used in this embodiment is as follows: Figure 2 As shown, the dimensions are 150mm×180mm. The material has two types of defects pre-fabricated: close-fitting and perforated. Each type has three defects of different sizes.

[0036] The sample 1 is placed on an optical platform with the detection surface of the sample 1 perpendicular to the field of view of the camera 7. At the same time, the speckle interferometer 9 and infrared equipment are arranged to make the sample 1 fill the field of view as much as possible. Figure 3 A schematic diagram of the optical path for measuring rubber-metal bonded structures using a laser shear speckle interferometer 9 is provided. The optical path includes a laser 2, a normally placed first reflecting mirror 4, a slightly tilted second reflecting mirror 5, a beam splitter 3, and a phase shifter 6. Light rays reflected from two closely spaced points on the surface of sample 1 pass through the first reflecting mirror 4 and the second reflecting mirror 5, respectively, and are focused onto the same point on the target surface of the camera 7, forming coherence. Therefore, two misaligned and overlapping images can be observed on the imaging plane.

[0037] like Figure 4 The diagram shows the experimental setup and arrangement. Sample 1 is placed on an optical platform. The speckle interferometer 9 and infrared equipment are placed parallel to the surface of sample 1. The infrared detector is placed adjacent to the speckle interferometer 9. (Not shown in the diagram.)

[0038] A laser speckle detection image enhancement method that fuses infrared images, such as Figure 1 As shown, the specific operation steps are as follows:

[0039] (1) Place the sample on the optical platform with the detection surface of the sample perpendicular to the camera field of view. At the same time, set up the speckle interferometer and infrared equipment to make the sample fill the field of view as much as possible. Adjust the shearing mirror to a suitable shearing angle so that the two overlapping images on the imaging surface present a certain amount of misalignment. The initial shearing angle can be set to 45°. Measure the deformation generated and register the speckle image and infrared image by methods such as transmission transformation, affine transformation or monochromatic matrix.

[0040] (2) During the experiment, the sample was loaded by heating with a halogen lamp to cause it to debond or the local area rich in glue to undergo relative deformation. The surface temperature change of the object was monitored in real time using an infrared device to obtain the location of the defect and extract the geometric center of the defect. At the same time, a speckle sequence image was collected by a speckle interferometer.

[0041] (3) Using the Shear+AI software developed by the inventors, the obtained speckle deformation map of the sample is automatically filtered and dephased to obtain the phase change value caused by the defect location of the sample, and a sequence phase map is acquired, such as... Figure 5 As shown;

[0042] The actual deformation caused by the defect location in the sample can be obtained from the formula for the optical path difference based on the phase change value. The relationship between the phase change value and the actual deformation value is as follows:

[0043]

[0044] In the formula, w represents the relative deformation generated by the two detection surfaces, and λ represents the wavelength of the laser. This represents the phase change value;

[0045] (4) Extract the region of interest from the phase map of the speckle sequence acquired in step (3) and perform a one-dimensional Fourier transform preprocessing operation to obtain the phase map at the first frequency, so as to further remove background noise. Figure 6 As shown; the expression for transforming the image from the time domain to the frequency domain, that is, transforming f(x) into F(ω), is:

[0046]

[0047] (5) Perform a series of morphological operations such as filtering and image segmentation on the phase map obtained in step (4) to obtain the binary map of the defect and the boundary of the defect, such as... Figure 7 As shown;

[0048] (5.1) Perform median filtering on the phase map at the first frequency obtained after one-dimensional Fourier transform to further filter out the speckle noise in the speckle phase map.

[0049] (5.2) The filtered phase map is segmented using a fixed threshold to obtain a binary map. Then, the binary map is opened to eliminate noise and small connected components in the image. Then, the area reduction of larger connected components caused by erosion is compensated by dilation operation. After this operation, the approximate boundary of the defect can be obtained.

[0050] (6) Taking advantage of the fact that the phase of the speckle phase map is all in (-π,π), the phase value at a certain position changes cyclically between (-π,π). When the phase value at that position changes abruptly, a new first-order fringe is generated. Track the gray value change of a pixel near the geometric center, and use the pixel difference between two adjacent moments to represent the derivative of the pixel gray value function. When the derivative changes abruptly, a new first-order fringe is generated. During the measurement process, the speckle fringes will continuously generate new first-order fringes with deformation. The number of times the derivative changes abruptly is used as the threshold. By setting an appropriate threshold, the phase map of the current defect when it is in the low-order fringe can be extracted by the position and boundary of the defect geometric center.

[0051] (8) By fusing the low-level fringe patterns of all defects into a single phase image, a clear and high-quality phase image can be obtained in the same image, achieving the purpose of image enhancement, such as... Figure 8 As shown.

Claims

1. A laser speckle detection image enhancement method that fuses infrared images, characterized in that, Includes the following steps: (1) Place the sample on the optical platform with the detection surface of the sample perpendicular to the camera field of view. At the same time, set up the speckle interferometer and infrared equipment to make the sample fill the field of view. Adjust the shearing mirror to a suitable shearing angle so that the two overlapping images on the imaging surface show a misalignment. The initial shearing angle is set to 45°. Measure the deformation produced. Register the speckle image and the infrared image by transmission transformation, affine transformation or monochromatic matrix method. (2) Apply load to the sample by heat loading to cause it to debond or the local area rich in glue to undergo relative deformation. Use infrared equipment to monitor the surface temperature change of the object in real time, obtain the location of the defect and extract the geometric center of the defect, and record the deformation diagram by speckle interferometer. (3) The obtained speckle deformation map of the sample is automatically filtered and dephased to obtain the phase change value generated by the defect location of the sample, and the speckle sequence phase map is collected. The actual deformation caused by the defect location in the sample is obtained from the formula for the optical path difference based on the phase change value, where the relationship between the phase change value and the actual deformation value is: In the formula The relative deformation generated by the two detection surfaces is calculated. The wavelength of the laser. This represents the phase change value; (4) Extract the region of interest from the phase map of the speckle sequence acquired in step (3) and perform one-dimensional Fourier transform preprocessing to obtain the phase map at the first frequency in order to further remove background noise. The image is transformed from the time domain to the frequency domain, that is... Transform into The expression is: ； (5) Filter and segment the phase image obtained in step (4) to obtain the binary image of the defect and the boundary of the defect. Specifically, this includes the following steps: (5.1) Perform median filtering on the phase map at the first frequency obtained after one-dimensional Fourier transform to further filter out the speckle noise in the speckle phase map; (5.2) The filtered phase map is segmented using a fixed threshold to obtain a binary map. Then, the binary map is opened to eliminate noise and small connected components in the image. Then, the area reduction of larger connected components caused by erosion is compensated by dilation operation. After this operation, the boundary of the defect is obtained. (6) Using the speckle phase diagram, the enclosed phases are all The characteristic is that the phase value at a certain position is The phase values ​​change cyclically between these points. When a sudden change occurs in the phase value at a given location, a new first-level fringe is generated. The gray value changes of a pixel near the geometric center are tracked. The derivative of the pixel gray value function is represented by the pixel difference between two adjacent moments. When a sudden change occurs in the derivative, a new first-level fringe is generated. The number of times the derivative changes abruptly is used as a threshold. By setting an appropriate threshold, the phase map of the current defect when it is in a low-level fringe can be extracted based on the location and boundary of the defect's geometric center. (7) By fusing the low-level stripe patterns of all defects into a single phase image, a clear and high-quality phase image can be obtained in the same image, thus achieving the purpose of image enhancement.

Images