Portable all-in-one machine integrating digital imaging and multi-mode film viewing of industrial nondestructive testing film and method

Abstract

The invention discloses a portable all-in-one machine integrating digital imaging and multi-mode film viewing of an industrial nondestructive testing film and a method, and belongs to the field of industrial radiography nondestructive testing. According to the portable all-in-one machine provided by the invention, the imaging module is used for shooting the industrial nondestructive testing film, the dynamic range and the signal-to-noise ratio of the image are improved by a multi-exposure method and a high-dynamic image processing synthesis algorithm, and high-resolution, high-dynamic and high-quality digital imaging of the industrial film is realized. Meanwhile, the portable all-in-one machine has the advantages of being smaller in size and convenient to operate and carry, an acceleration coprocessor is arranged in the portable all-in-one machine to assist in detection, defect prompts can be rapidly given out without networking, the multi-mode film watching function of simulating film watching, digital picture watching and virtual reality film watching is provided for engineers on site, and the efficiency is improved. Therefore, rapid detection of the industrial welding quality is realized, and the efficiency and the performance of industrial nondestructive detection are improved.

Description

Technical Field

[0001] This invention belongs to the field of industrial radiographic nondestructive testing, specifically relating to a defect detection method and portable all-in-one device that integrates digital imaging and multi-mode viewing of industrial nondestructive testing films. Background Technology

[0002] Industrial non-destructive testing (NDT) typically involves engineers manually inspecting X-ray films to detect defects. This method suffers from drawbacks such as low efficiency, high labor intensity, system complexity, and low accuracy. However, with the development of electronic, computer, and artificial intelligence technologies, the digitization of industrial films has provided the conditions for information-based management and intelligent processing of defect detection, and is receiving increasing attention.

[0003] Currently, the primary method for digitizing industrial X-ray film is to use CCD scanners to scan the film line by line to obtain high-resolution images. Common scanners for industrial X-ray film include the NDT Pro from Vidar (USA), the FS50B from Waygate (USA), and the 2905 / HD from Array (Japan). These scanners require a desktop computer for scanning and feature high resolution, good image quality, and a wide dynamic range. However, scanning a single frame typically takes tens of seconds, resulting in slow imaging speeds and high equipment costs. While some companies use area array imaging sensors to capture images of the film, achieving speeds of less than one second, this method usually has low resolution and is limited by the sensor, resulting in a low dynamic range that cannot meet the image quality requirements for defect detection.

[0004] CN216956532U, "An LED Industrial X-ray Film Viewing Lamp with Adjustable Clamping Structure," discloses an LED industrial X-ray film viewing lamp with an adjustable clamping structure. This lamp can be used for manual observation of industrial non-destructive testing films and has a simulated viewing function. However, it suffers from strong light leakage and glare during viewing, and it cannot perform digital imaging of the film, making information management inconvenient. CN218157633U, "An Integrated Film Digital Evaluation Machine," discloses a scanning film digitization system. CN117714838A, "An Industrial X-ray Film Digital Imaging Device," discloses a device and method for batch scanning films using a motor. However, these systems only support film digitization, have slow scanning speeds, or have image quality limited by sensor performance. They generally require integration with other devices such as computers for digital viewing to achieve digital image display and intelligent processing. Furthermore, these systems do not support on-site manual simulated viewing, cannot quickly provide defect results, and are inconvenient for defect verification. CN108665452B, "Method and System for Scanning and Entering Pipeline Weld Films into a Database and Identifying Weld Defects," discloses a system that first scans and enters the film into a database and then performs intelligent processing. CN115661057A, "An Industrial Non-destructive Testing System and Method Based on Cloud-Edge Collaboration and Deep Learning," discloses an industrial defect detection system that combines a field terminal and a cloud data processing center. The systems disclosed in the above patent documents do not support on-site manual viewing of the films and are relatively complex.

[0005] In summary, the technologies involved in the current patent documents have shortcomings such as low imaging quality, separation of imaging and viewing, incomplete functions, complex and bulky systems, and the need for network connectivity. They do not support portability and do not integrate film viewing and digital imaging into a single portable device. Therefore, they cannot be well adapted to scenarios that require rapid response, such as small-scale industrial welding projects, projects in narrow construction spaces, and partial rework projects. Summary of the Invention

[0006] The purpose of this invention is to provide a portable integrated device and method for industrial film digitization and viewing, which can simplify the operation process of industrial film viewing and digitization, and improve efficiency and portability.

[0007] To achieve the above objectives, this invention first provides a portable all-in-one device integrating digital imaging and viewing of industrial non-destructive testing film. This device uses an imaging module to capture images of the film, employing a multi-exposure method and a high dynamic range image processing synthesis algorithm to improve the dynamic range and signal-to-noise ratio of the image, achieving high-resolution, high-dynamic, and high-quality digital imaging of industrial film. Simultaneously, this invention utilizes an integrated device, featuring smaller size, ease of operation and portability. Furthermore, the device incorporates a built-in accelerated coprocessor, enabling rapid auxiliary defect detection. It can quickly provide defect alerts for the current film without requiring a network connection, and offers engineers multi-mode viewing functions including simulated film viewing, digital image viewing, and virtual reality viewing. This enables rapid inspection of industrial welding quality and improves the efficiency and performance of industrial non-destructive testing.

[0008] In a first aspect of the invention, a portable all-in-one machine for industrial non-destructive testing is provided, integrating the functions of digital imaging and multi-mode viewing of industrial non-destructive testing films. It includes a processing host module, an imaging module, a backlight module, a touch display module, a storage module, a wireless communication module, and a power supply module. The processing host module is connected to the imaging module, backlight module, touch display module, storage module, and wireless communication module. The power supply module provides power to the aforementioned modules. All modules are mounted and fixed in a portable chassis. The backlight module includes a programmable LED array, where each LED can be individually controlled for switching and brightness. The industrial non-destructive testing film is placed on the backlight module. The imaging module includes… A programmable camera with controllable focal length, aperture, and imaging angle is used to capture images of industrial non-destructive testing film on a backlight module and measure ambient light intensity. The processing host module integrates a high-performance processor, an accelerator coprocessor, and peripheral expansion interfaces to control the illumination range and brightness of the backlight module and synthesize multiple film images captured by the imaging module under different backlight conditions into a high dynamic range image. The touch display module includes a touch screen for displaying the high dynamic range image and acquiring user touch input, marking defect information on the high dynamic range image. The storage module stores the high dynamic range image and defect information data. The wireless communication module communicates with a remote server to upload image data and perform remote verification.

[0009] In addition, this portable all-in-one machine can also connect to an external mouse and keyboard via the interface set on the host module.

[0010] Furthermore, the aforementioned portable all-in-one machine also includes an audio input module and an audio output module connected to the processing host module. The audio input module is used to collect the operator's voice input to realize voice-to-text input and voice control of the all-in-one machine. The audio output module is used to play the audio output of the all-in-one machine and provide voice prompts.

[0011] Furthermore, the processing host module incorporates a high-performance processor, an accelerator coprocessor, and peripheral expansion interfaces. It is used to: determine the illumination range of the backlight module based on user input and ambient light intensity; control the backlight module to perform multi-level backlight control; simultaneously send shooting commands and parameters to the imaging module; capture images of the industrial non-destructive testing film on the backlight module; acquire multiple film images under different backlight conditions; execute high dynamic range image processing and synthesis algorithm software to complete high dynamic range imaging of the industrial film; and store the high dynamic range image photos in the storage module; display the high dynamic range image photos on the touch display module; obtain user touch input from the touch display module; annotate defect information such as defect location and type on the high dynamic range image and store it together with the high dynamic range image; and support transmitting the high dynamic range image and defect information to a remote server via a wireless communication module for data backup and remote verification.

[0012] Furthermore, the imaging module includes a camera with programmable control over focal length, aperture, and imaging angle. It is used to: capture images of industrial non-destructive testing film; measure the intensity of ambient light based on the film images to determine the illumination range of the backlight module; determine the position of the film on the backlight module based on the film images to determine the backlight module's emission range and avoid backlight leakage and glare; and acquire multiple images under different backlight intensities for high dynamic range image synthesis. The focal length of the imaging module is adjustable, enabling automatic focusing of the film and image capture at different focal lengths. Preferably, the imaging module incorporates a rotatable reflector; by controlling the rotation angle of the reflector, different parts of the industrial film can be captured, expanding the shooting range.

[0013] Furthermore, the backlight module consists of an array of high-brightness LEDs (see...). Figure 4 Each LED can be individually controlled for its on / off state and light intensity. The backlight module has a uniform light distribution layer covering the LED array, providing even backlight illumination for the industrial film placed on top of it. The backlight module is used to: provide backlight for the industrial film; the backlight range can be adjusted according to the film size and position; prevent direct backlight from shining into the eyes of the viewer; and the backlight intensity can be varied to provide illumination conditions for capturing images under different backlight conditions.

[0014] Furthermore, the touch display module includes a high-resolution touch screen that supports high-resolution and high-dynamic-range image display and multi-touch functionality. It is used to: display high-dynamic-range image photographs of industrial films; allow viewers to annotate defect information such as defect location, type, and grade on the screen; support digital viewing; and support operations such as data saving, data retrieval, and system settings.

[0015] Furthermore, the wireless communication module communicates with a remote server via a WiFi local area network or a mobile communication network to enable image data uploading and remote verification.

[0016] Furthermore, the power module incorporates an AC-DC converter circuit, a DC-DC converter circuit, and a battery charging and discharging circuit to provide power for the entire portable all-in-one device. It supports three power supply modes: external 220V AC power input, external (12V~28V) DC power supply, and internal battery (12V~28V), providing power for the aforementioned modules.

[0017] Furthermore, the storage module has a built-in high-capacity solid-state drive for storing high dynamic range image data synthesized by the imaging module, digital viewing defect information data, device operation logs, etc.

[0018] Furthermore, the aforementioned portable all-in-one machine also includes an audio input module and an audio output module, both connected to the processing host module. The audio input module is used to collect the operator's voice input, enabling voice-to-text input and voice control of the all-in-one machine system. The audio output module is used to play the audio output of the all-in-one machine system, providing system voice prompts.

[0019] Furthermore, the imaging module is located at the top of the portable chassis, and the backlight module is located at the bottom of the portable chassis. The imaging module can have three imaging structures:

[0020] One of them is like Figure 6 The vertical imaging structure shown has the camera lens pointing directly downwards at the backlight module, allowing for planar imaging of the entire film on the backlight module.

[0021] Secondly, as Figure 7 The horizontal imaging structure shown involves the camera positioned on the left or right side of the upper part of the camera housing, facing horizontally to the right or left. A rotatable mirror is placed to the right or left of the lens for reflective imaging. By rotating the mirror, the horizontal range of the film imaging can be expanded, and overlapping imaging of the same area can also be achieved. If a linear camera is used, horizontal scanning imaging of the film can be achieved. This structure also has the advantage of reducing the vertical height of the camera housing.

[0022] Thirdly, as Figure 7 The illustrated horizontal imaging structure features a camera positioned at the rear or front of the upper part of the camera housing, with the camera lens facing horizontally forward or backward. A rotatable mirror is placed in front of or behind the lens for reflective imaging. By rotating the mirror, the front-to-back range of the film imaging can be expanded, and overlapping imaging of the same area can be achieved. If a linear camera is used, front-to-back scanning imaging of the film can be realized. This structure also has the advantage of reducing the vertical height of the camera housing.

[0023] This invention also provides a method for industrial defect detection using the aforementioned portable all-in-one machine integrating digital imaging and multi-mode viewing of industrial non-destructive testing films, comprising:

[0024] Step 1: The host module controls the imaging module to continuously take pictures of the film placement area on the backlight module and performs object detection on the acquired images. Simultaneously, the ambient light intensity is obtained by measuring the brightness of the images.

[0025] Step 2: Place a piece of industrial non-destructive testing film to be inspected onto the backlight module. The processing host module obtains the film's coordinates on the backlight module through the film image and controls the backlight module to turn on the LEDs under the film to provide backlight for film observation and imaging, while the LEDs located outside the film remain off. Figure 5 As shown.

[0026] Step 3: The host module controls the backlight area of ​​the backlight module to provide backlight of different intensities, and at the same time controls the camera module to shoot the film to obtain a multi-exposure image sequence under different backlight intensities.

[0027] Step 4: Based on the multi-exposure image sequence, the processing host module synthesizes a high dynamic range image photo with the help of the accelerated coprocessor in the processing host module, and stores the high dynamic range image photo in the storage module.

[0028] Step 5: The host module obtains a backlight intensity suitable for manual observation based on the high dynamic range image and multi-exposure image sequence, and sets the backlight module to that backlight intensity.

[0029] Step 6: The host module uses the ambient light intensity obtained earlier to correct the synthesized high dynamic range image, obtains the corrected high dynamic range image, and displays it on the touch screen.

[0030] Step 7: Users view the film in multiple modes to locate and detect defects in industrial film.

[0031] Manual simulation viewing mode: Engineers directly observe the film on the backlight module and use touch to mark the location, type and level of defects on the touch display module.

[0032] Manual digital viewing mode: Engineers can directly observe high dynamic range images displayed on the touch screen, and at the same time use the image magnification, filtering, grayscale adjustment and other tools provided by the processing host module to observe details, and mark the location, type and level of defects on the touch screen.

[0033] Intelligent Assisted Digital Viewing Mode: The processing host module inputs high dynamic range (HMR) images into an intelligent defect detection algorithm (such as the Mask RCNN defect detection algorithm trained with fine-tuning of defect samples). The algorithm outputs prompts about the location, type, and level of defects, which are then superimposed onto the HMR images on the touch screen to assist engineers in defect localization and detection.

[0034] Intelligent Assisted Virtual Reality Viewing Mode: A laser projection device is placed next to the camera in the imaging module. This device can directly project the defect location information output by the intelligent defect detection algorithm onto the film, guiding engineers to further observe the defects and achieve rapid defect location and detection.

[0035] When watching the footage, engineers can combine the above modes to achieve rapid defect location and defect detection.

[0036] Step 8: The processing host module will combine the defect detection results from the multi-mode film viewing, including defect location, defect type and defect level, with engineering management information (engineer information, location information, equipment information, time information, etc.), multi-exposure image sequences and high dynamic range image photographs to generate a film viewing record, and store it in the local storage module.

[0037] Step 9: The host module uploads the film viewing record to the remote server via the wireless communication module.

[0038] Step 10: This completes the digital imaging and multi-mode viewing of the industrial film. The engineer replaces the industrial non-destructive testing film onto the backlight module and returns to Step 2 above to repeat the operation.

[0039] Specifically, in steps 1 and 2 above, the host module uses an object detection algorithm to detect objects in the image. A calibrated YOLO object detection algorithm, such as YOLOV5, can be used to detect and locate industrial film, and output the four corner coordinates of the industrial film on the backlight module.

[0040] Specifically, the ambient light intensity in step 1 above is obtained by averaging the pixel values ​​outside the industrial film area.

[0041] Specifically, in step 4 above, the host processing module uses a high dynamic range image processing synthesis algorithm to infer the camera's response function from a sequence of multiple exposures, thereby synthesizing an HDR image. The specific steps are as follows:

[0042] Step 1: Define the camera response function:

[0043] That is, a function of pixel value, exposure time, and corresponding scene irradiance, defined as follows:

[0044] z ij =f(E i t j )

[0045] Among them, z ij E represents the pixel value of the i-th pixel in the j-th image. i It is the radiance received at that point, representing the true image information of that point, t j This is the exposure time of the j-th image.

[0046] Step 2: After inverting the left and right sides of the camera response function and then taking its logarithm, we have:

[0047] log2f -1 (z ij ) = log2E i +log2t j

[0048] Let the part on the left be g, then we have:

[0049] g(z ij ) = log2E i +log2t j

[0050] By finding the function g, the original camera response function can be obtained.

[0051] Step 3: Establish the objective function for optimization.

[0052] First, fix g(128) = 0, and transform the solution of the camera response function into a constrained extremum problem. That is, to find the illuminance E at each point. i This minimizes the value of the following error function:

[0053]

[0054] Where w is the weighting function, we consider pixels near 128 to have the best image quality, while pixels near 0 and 255 have poorer image quality. Therefore, we assign different weights to different pixels:

[0055]

[0056]

[0057] The second term, λ, is multiplied by the second-order smoothing term as a parameter, making the fitted function smoother.

[0058] Step 4: Obtain g(z) using singular value decomposition.

[0059] Due to pixel value z ijIt is discrete; for the 8-bit case, there are only 256 possible values. Therefore, singular value decomposition can be used to obtain the values ​​of g(z) in these 256 cases, thus obtaining the camera response function.

[0060] Step 5: Reconstruct the high dynamic range illuminance map {Ei}

[0061] After obtaining the function g, we can calculate the corresponding illuminance value E for each point using the following formula. i :

[0062] log2E i =g(z) ij )-log2t j

[0063] To be more robust and to effectively utilize information from all image sequences, we should use the pixel values ​​of all images at the same point and calculate the illuminance at that point using a weighted average.

[0064]

[0065] Thus, we can obtain the illuminance map {Ei} for all pixel locations, since E i It represents the true illumination information of the image at that point, containing information from multiple images at that point; {Ei} is also the synthesized high dynamic range image data.

[0066] Specifically, in step 5 above, the processing host module obtains the backlight intensity suitable for manual viewing by using a manual viewing backlight calculation algorithm. The specific method is to compare the multi-exposure image sequence with the synthesized high dynamic range image one by one, select the exposure image with the highest correlation, and output the corresponding backlight intensity as the backlight intensity for manual viewing.

[0067] Specifically, in step 6 above, the host module uses a high dynamic range image display ambient light correction algorithm to correct the high dynamic range image photograph synthesized in step 4. The specific method is as follows: the high dynamic range image data is tone-mapped to the pixel value range of [0, 255], and then adaptive contrast enhancement is performed. During adaptive contrast enhancement, an ambient light intensity factor h is added to the enhancement factor P. When the ambient light is strong, the degree of adaptive contrast enhancement is increased, making the image contrast stronger and facilitating the observation of details. Finally, the image after the above contrast enhancement is displayed on the touch display module's screen for engineers to perform multi-mode viewing. The specific steps of the above adaptive contrast enhancement are as follows:

[0068] Step 1: Calculate the global luminance mean square error σ for the entire image, and determine the contrast enhancement factor P0 using the following method:

[0069] If σ≤3, then P0=3.

[0070] If σ≥10, P0=1,

[0071] If 3 < σ < 10, then P0 takes linear interpolation, P0 = (27 - 2σ) / 7;

[0072] Step 2: Normalize the ambient light intensity to [0,1] to obtain the ambient light intensity factor h. Take P = P0 + K * h, where K is the ambient light intensity amplification parameter. K takes a value between 1 and 3, with 1.5 as the default value.

[0073] Step 3: For each pixel I(X,Y), calculate the Gaussian convolution G(X,Y) of that pixel, and calculate its exponent E(X,Y) = (G(X,Y) / I(X,Y)). P ;

[0074] Step 4: Calculate the adaptively enhanced pixel value S(X,Y) = 255 * In(X,Y) for each pixel. E(X，Y) , where In(X,Y) is the normalized value of the original image pixel I(X,Y): In(X,Y)=I(X,Y) / 255.

[0075] Compared with existing technologies, the portable industrial film digital imaging and viewing integrated machine and industrial defect detection method of the present invention have the following main advantages:

[0076] 1. Excellent portability, allowing for film viewing and digitization anytime, anywhere. Existing film digitization instruments typically require a desktop computer, making them bulky and inconvenient to carry. The all-in-one device of this invention is smaller and easier to operate and carry.

[0077] 2. High speed: While maintaining image quality comparable to existing film digitization instruments, it allows for faster film viewing and digitization. This is because existing instruments typically use scanning imaging methods, which are relatively slow. The imaging module in this invention, when using an area array imaging sensor, can complete imaging much faster. Simultaneously, it employs a high dynamic range image synthesis algorithm, accelerated by a built-in accelerator coprocessor, providing rapid imaging capabilities while ensuring image quality.

[0078] 3. High imaging quality: Compared with existing film digitization instruments that use photographic imaging, the multiple exposure synthesis method used in this invention can improve the dynamic range and signal-to-noise ratio of the image, thereby improving the imaging quality. The quality reaches or exceeds that of film digitization instruments that use scanning imaging. When the imaging module is configured with a reflective mirror, the imaging range can be increased. More details can also be obtained through the overlap of imaging ranges, generating higher quality high dynamic range images.

[0079] 4. High detection efficiency: This invention combines multiple viewing methods, including manual simulation viewing, digital viewing, intelligent assisted viewing, and virtual reality viewing, which can reduce the workload of engineers and improve detection accuracy and efficiency. Attached Figure Description

[0080] Figure 1 This is a three-dimensional schematic diagram of the internal components of the portable industrial film digital imaging and viewing integrated device described in the embodiment.

[0081] Figure 2 yes Figure 1 The right view of the interior of the portable industrial film digitization imaging and viewing device shown.

[0082] Figure 3 This is a block diagram of the portable industrial film digital imaging and viewing integrated device described in the embodiment.

[0083] Figure 4 This is a schematic diagram of the programmable backlight module described in the embodiment (the top light-diffusing layer has been removed).

[0084] Figure 5 This is a schematic diagram of the programmable backlight module in operation when the industrial film is placed (the backlight under the film is turned on, the backlight LED outside the film is turned off, and the top light-diffusing layer has been removed).

[0085] Figure 6 This is a schematic diagram of the internal structure of the vertical imaging module of the programmable imaging module.

[0086] Figure 7 This is a schematic diagram of the internal structure of the horizontal imaging module, which is positioned horizontally on the left and right sides.

[0087] In the diagram: 1-Processing host module, 2-Programmable imaging module, 3-Programmable backlight module, 4-Touch display module, 5-Power supply module, 21-Camera, 22-Lens, 23-Reflector, 24-Rotating support rod. Detailed Implementation

[0088] The technical solution of the present invention will be clearly and completely described below with reference to the accompanying drawings and embodiments, but this does not limit the scope of the present invention.

[0089] like Figure 1 , Figure 2 and Figure 3As shown, the portable industrial film digitization imaging and viewing all-in-one device of this embodiment includes: a processing host module 1, a programmable imaging module 2, a programmable backlight module 3, a touch display module 4, a large-capacity storage module, an audio input module, an audio output module, a wireless communication module, and a power module 5. The processing host module 1 is connected to the programmable imaging module 2, the programmable backlight module 3, the touch display module 4, the storage module, the audio input module, the audio output module, and the wireless communication module. The power module 5 supplies power to the above modules, and all modules are installed and fixed in a portable chassis. In addition, this all-in-one device can also be connected to an external mouse and keyboard.

[0090] The processing host module 1 is located on the top of the all-in-one device. Figure 1 The processing host module 1 is equipped with a built-in processor, an accelerator coprocessor, and peripheral expansion interfaces. It uses an RK3588 embedded host, providing HDMI, USB, and M.2 PCIe interfaces. The functions of the processing host module 1 include: determining the illumination range of the programmable backlight module 3 based on user input and ambient light intensity; controlling the programmable backlight module 3 for multi-level backlight control; sending shooting commands and parameters to the programmable imaging module 2; capturing images of the industrial film on the programmable backlight module 3; acquiring multiple film images under different backlight conditions; executing high dynamic range image processing and synthesis algorithms to complete high dynamic range imaging of the industrial film; storing the high dynamic range images in a storage module; displaying the high dynamic range images on a touch display module 4; acquiring user touch input from the touch display module 4; marking defect information such as defect location and type on the image and storing it along with the high dynamic range images; and supporting the transmission of the high dynamic range images and defect information to a remote server via a wireless communication module for backup and remote verification.

[0091] The programmable imaging module 2 is located on the top of the integrated device. Figure 1 This includes a camera with programmable control over focal length, aperture, and imaging angle. The camera uses an IMX586 image sensor and connects to the processing host module via a USB or MIPI-CSI interface. The programmable imaging module 2 functions to: capture images of industrial film; measure ambient light intensity based on the film images to determine the illumination range of the programmable backlight module 3; determine the film's position on the backlight based on the film images to determine the illumination range of the programmable backlight module 3, avoiding backlight leakage and glare; and acquire multiple images under different backlight intensities for high dynamic range image synthesis. The focal length of the programmable imaging module 2 is adjustable, enabling automatic focusing of the film and image capture at different focal lengths. The programmable imaging module 2 also includes a built-in rotatable reflector. Figure 7 By controlling the rotation angle of the reflector 23, it is possible to shoot different parts of the industrial film and expand the shooting direction.

[0092] The programmable backlight module 3 is located at the bottom of the all-in-one device. Figure 1 ),like Figure 4 As shown, the backlight module 3 is composed of an array of high-brightness LEDs, each of which can be individually controlled for switching on and off and its luminous intensity. The top layer of the programmable backlight module 3 is covered with a light-diffusing layer, providing uniform backlight illumination for the industrial film placed on top of it. The programmable backlight module 3 connects to the processing host module via an RS485 interface and a USB-RS485 interface. The functions of the programmable backlight module 3 include providing backlight for the industrial film, with the backlight range adjustable according to the size and position of the film. Figure 5 As shown, this design prevents the backlight from shining directly into the eyes of the operator, while also allowing the backlight intensity to vary to provide illumination conditions for acquiring images under different backlight conditions.

[0093] The touch display module 4 is disposed on the side wall of the all-in-one device. Figure 1 The system includes a high-resolution 15.6-inch touchscreen display and its controller, with a display resolution of 3840×2160. The display is connected to the processing host module 1 via an HDMI interface, supporting high-resolution and high dynamic range image display. The touchscreen controller is connected to the processing host module 1 via a USB interface, supporting multi-touch functionality. The touchscreen module's functions include: displaying high dynamic range images of industrial films; allowing viewers to annotate defect information such as location, type, and grade on the screen; supporting digital viewing; and supporting data saving, data retrieval, and system settings operations.

[0094] The audio input module is connected to the processing host module 1 via an SPI interface and is used to collect the operator's voice input to realize voice and text information input and system voice control.

[0095] The audio output module is connected to the HDMI interface of the processing host module 1 and is used to play the system's audio output and provide the system's voice prompt function.

[0096] The wireless communication module communicates with a remote server via a built-in WiFi module or a mobile communication network connected through a USB-extended 4G mobile communication module, enabling image data uploading and remote verification.

[0097] The power module 5 is located on the top of the all-in-one device. Figure 1It has built-in AC-DC conversion circuit, DC-DC conversion circuit and battery charging and discharging circuit to provide power to the entire device. It supports three power supply modes: external 220V AC power input, external (12V~28V) DC power supply and internal battery (12V~28V) to provide power to the above modules.

[0098] The storage module is equipped with a large-capacity NVMe SSD of specification 2242 through the M.2 PCIe interface of the host module 1. It is used to store high dynamic range image data synthesized by the programmable imaging module 2, digital viewing defect information data, device operation logs, etc.

[0099] The entire device uses an aluminum alloy frame to assemble the various modules, forming a portable industrial film digital imaging and multi-mode film viewing integrated device, such as... Figure 1 As shown.

[0100] The operation method of the aforementioned portable industrial film digital imaging and multi-mode film viewing integrated device includes:

[0101] Step 1: After powering on, if Figure 1 After completing its system self-test, the integrated device shown uses the processing host module 1 to control the programmable imaging module 2 to continuously take pictures of the film placement area on the programmable backlight module 3, and then performs object detection on the acquired images. Simultaneously, the ambient light intensity of the integrated device is obtained by measuring the brightness of the images.

[0102] Step 2: When the engineer places an industrial non-destructive testing film to be inspected onto the programmable backlight module 3, the processing host module 1 runs the object detection algorithm to obtain the placement coordinates of the film on the programmable backlight module 3. It then controls the programmable backlight module 3 to turn on the LEDs under the film to provide backlight for film observation and imaging, while the LEDs located outside the film remain off. Figure 5 As shown.

[0103] Step 3: The host module 1 controls the backlight area of ​​the programmable backlight module 3 to provide backlight of different intensities, and at the same time controls the programmable imaging module 2 to take pictures of the film to obtain a multi-exposure image sequence under different backlight intensities.

[0104] Step 4: The host module 1 inputs the multi-exposure image sequence into a high dynamic range image processing and synthesis algorithm software. With the help of the accelerated coprocessor in the host module 1, a high dynamic range image is synthesized and stored in the storage module.

[0105] Step 5: The host module 1 inputs the high dynamic range image and multi-exposure image sequence into the manual viewing backlight calculation algorithm software, outputting a backlight intensity suitable for manual observation. The programmable backlight module 3 is then set to this backlight intensity.

[0106] Step 6: The host module 1 processes the synthesized high dynamic range image and the ambient light intensity obtained earlier into a high dynamic range image display ambient light correction algorithm software, outputs the corrected high dynamic range image photo, and displays it on the touch display module 4.

[0107] Step 7: Users view the film in multiple modes to locate and detect defects in industrial film.

[0108] Manual simulation viewing mode: Engineers directly observe the film on the programmable backlight module 3 and use touch to mark the location, type and level of defects on the touch display module 4.

[0109] Manual digital viewing mode: Engineers can directly observe the high dynamic range images displayed on the touch screen, and at the same time use the image magnification, filtering, grayscale adjustment and other tools provided by the processing host module 1 to observe details, and mark the location, type and level of defects on the touch screen.

[0110] Intelligent Assisted Digital Viewing Mode: The host module 1 inputs high dynamic range (HMR) images into an intelligent defect detection algorithm software. The software outputs prompts about the location, type, and level of defects, which are then overlaid onto the HMR images on the touch screen to assist engineers in further defect localization and detection.

[0111] Intelligent Assisted Virtual Reality Viewing Mode: A laser projection device is set up next to the camera in the programmable imaging module 2. This device can directly project the defect location information output by the intelligent defect detection algorithm software onto the film, guiding engineers to further observe the defects and achieve rapid defect location and defect detection.

[0112] When watching the footage, engineers can combine the above modes to achieve rapid defect location and defect detection.

[0113] Step 8: The processing host module 1 takes the defect detection results from the multi-mode film viewing, including defect location, defect type and defect level, and combines them with engineering management information (engineer information, location information, equipment information, time information, etc.), multi-exposure image sequences and high dynamic range image photos to generate a film viewing record, and stores it in the local storage module.

[0114] Step 9: The host module 1 uploads the film viewing record to the remote server via the wireless communication module.

[0115] Step 10: The digital imaging and multi-mode viewing of the industrial film are now complete. The engineer replaces the industrial film with a new one on the programmable backlight module 3 and returns to step 2 to repeat the operation.

[0116] The above description of the embodiments is merely for the purpose of helping to understand the method and core ideas of the present invention, and is intended to enable those skilled in the art to understand the content of the present invention and implement it accordingly. It should not be construed as limiting the scope of protection of the present invention. All equivalent modifications or substitutions made without departing from the spirit of the present invention are included within the scope defined by the claims of the present invention.

Claims

1. A portable all-in-one machine for industrial non-destructive testing, characterized in that, The system includes a processing host module, an imaging module, a backlight module, a touch display module, a storage module, a wireless communication module, and a power supply module. The processing host module is connected to the imaging module, backlight module, touch display module, storage module, and wireless communication module. The power supply module provides power to all these modules. All modules are mounted and fixed in a portable chassis. The backlight module includes a programmable LED array, where each LED can be individually controlled for switching on / off and brightness. Industrial non-destructive testing film is placed on the backlight module. The imaging module includes a camera with programmable control over focal length, aperture, and imaging angle for capturing images of the backlight module. The system employs an industrial non-destructive testing film to acquire film images and measure ambient light intensity. The processing host module integrates a high-performance processor, an accelerator coprocessor, and peripheral expansion interfaces to control the illumination range and brightness of the backlight module and synthesize multiple film images captured by the imaging module under different backlight conditions into a high dynamic range image. The touch display module includes a touch screen for displaying the high dynamic range image and acquiring user touch input, marking defect information on the high dynamic range image. The storage module stores the high dynamic range image and defect information data. The wireless communication module communicates with a remote server to upload image data and perform remote verification.

2. The portable all-in-one machine as described in claim 1, characterized in that, The portable all-in-one device also includes an audio input module and an audio output module connected to the processing host module. The audio input module is used to collect the operator's voice input to realize voice-to-text input and voice control of the all-in-one device. The audio output module is used to play the audio output of the all-in-one device and provide voice prompts.

3. The portable all-in-one machine as described in claim 1, characterized in that, The processing host module is equipped with interfaces for connecting an external mouse and keyboard.

4. The portable all-in-one machine as described in claim 1, characterized in that, The imaging module is located at the top of the chassis, and the backlight module is located at the bottom of the chassis. The imaging module has one of the following imaging structures: First, a vertical imaging structure, in which the camera lens faces downwards and is aligned with the backlight module to perform planar imaging of the entire film on the backlight module; Second, a horizontal imaging structure, in which the camera is located on the left or right side of the top of the chassis, the camera lens is horizontally facing right or left, and a rotatable reflector is set on the right or left side of the lens for reflective imaging; Third, a horizontal imaging structure, in which the camera is located at the rear or front of the top of the chassis, the camera lens is horizontally facing forward or backward, and a rotatable reflector is set on the front or rear side of the lens for reflective imaging.

5. The portable all-in-one machine as described in claim 1, characterized in that, The backlight module has a light-diffusing layer covering the LED array.

6. An industrial defect detection method, characterized in that, Using the portable all-in-one machine described in any one of claims 1 to 5, mathematical imaging and multi-mode viewing of industrial non-destructive testing films are performed, achieving industrial defect detection through the following steps: 1) The processing host module controls the imaging module to continuously take pictures of the film placement area on the backlight module and perform object detection on the acquired images; at the same time, the ambient light intensity is obtained by measuring the brightness of the images. 2) Place an industrial non-destructive testing film onto the backlight module. The processing host module obtains the film's placement coordinates on the backlight module through the film image and controls the backlight module to turn on the LEDs under the film to provide backlight for film observation and imaging, while the LEDs located outside the film remain off. 3) The host module controls the backlight area of ​​the backlight module to provide backlight of different intensities, and at the same time controls the camera module to shoot the film to obtain a multi-exposure image sequence under different backlight intensities. 4) The processing host module synthesizes a high dynamic range image based on the multi-exposure image sequence and stores the high dynamic range image in the storage module; 5) The processing host module obtains a backlight intensity suitable for manual observation based on high dynamic range image photos and multi-exposure image sequences, and sets the backlight module to that backlight intensity; 6) The host module uses the ambient light intensity obtained in step 1) to correct the high dynamic range image photo synthesized in step 4), obtains the corrected high dynamic range image photo, and displays it on the touch screen. 7) Employ one or more of the following viewing modes for multi-mode viewing to achieve defect location and detection of industrial film: Manual simulated film viewing mode: Engineers directly observe the film on the backlight module and mark the location, type, and level of defects on the touch screen using touch. Manual digital viewing mode: Engineers directly observe the high dynamic range images displayed on the touch screen, while using the image processing tools provided by the processing host module for detailed observation, and marking the location, type and level of defects on the touch screen using touch. Intelligent Assisted Digital Viewing Mode: The processing host module obtains the location, type, and level of defects on high dynamic range (HMR) images through an intelligent defect detection algorithm, and overlays it onto the HMR images on the touch screen to assist engineers in defect location and detection. Intelligent Assisted Virtual Reality Viewing Mode: A laser projection device is placed next to the camera in the imaging module. The defect location information obtained by the intelligent defect detection algorithm is directly projected onto the film through the laser projection device, guiding engineers to observe the defects and achieving rapid defect location and defect detection. 8) The processing host module generates a film viewing record by combining the defect detection results of multi-mode film viewing with engineering management information, multi-exposure image sequences, and high dynamic range image photographs, and stores it in the storage module; 9) The processing host module uploads the film viewing records to the remote server through the wireless communication module.

7. The industrial defect detection method as described in claim 6, characterized in that, In steps 1) and 2), the host module uses an object detection algorithm to detect objects in the image. By detecting and locating the film used in industrial non-destructive testing, the coordinates of the four corners of the film on the backlight module are obtained. In step 1), the ambient light intensity is obtained by averaging the pixel values ​​outside the film area.

8. The industrial defect detection method as described in claim 6, characterized in that, In step 4), the host processing module uses a high dynamic range image processing and synthesis algorithm to infer the camera's response function from a multi-exposure image sequence, thereby synthesizing a high dynamic range image, including: 41) Define the camera response function, which is a function of pixel value, exposure time, and corresponding scene irradiance, as follows: z ij =f(E i t j ) Among them, z ij E represents the pixel value of the i-th pixel in the j-th image. i It represents the illuminance received at that point, indicating the true image information of that point. j This is the exposure time of the j-th image; 42) Taking the left and right inverses of the camera response function, and then taking its logarithm, we have: log2f -1 (With ij )=log2E i +log2t j Let the part on the left be g, then we have: g(z ij )=log2E i +log2t j Finding the function g yields the original camera response function; 43) Establish the optimization objective function First, fix g(128) = 0, and transform the solution of the camera response function into a constrained extremum problem, that is, to find the illuminance E at each point. i This minimizes the value of the following error function: Here, w is a weighting function, which considers pixels near 128 to have the best image quality, while pixels near 0 and 255 are considered to have poorer image quality. Therefore, different weights are assigned to different pixels: The second term, λ, is multiplied by the second-order smoothing term as a parameter, making the fitted function smoother. 44) Obtain g(z) using singular value decomposition. Due to pixel value z ij It is discrete; for the 8-bit case, there are only 256 possible values. Therefore, singular value decomposition is used to obtain the values ​​of g(z) in these 256 cases, which gives us the camera response function. 45) Reconstruct the high dynamic range illuminance map {Ei} After obtaining the function g, calculate the corresponding illuminance value E for each point using the following formula. i : log2E i =g(z ij )-log2t j The illumination at a given point is calculated by taking the weighted average of the pixel values ​​from all images at that point. In this way, the illumination map {Ei} for all pixel locations is obtained, which is the synthesized high dynamic range image data.

9. The industrial defect detection method as described in claim 6, characterized in that, In step 5), the host module obtains the backlight intensity suitable for manual viewing by using a manual viewing backlight calculation algorithm. Specifically, it compares the multi-exposure image sequence with the synthesized high dynamic range image one by one, selects the exposure image with the highest correlation, and uses the corresponding backlight intensity as the backlight intensity suitable for manual viewing.

10. The industrial defect detection method as described in claim 6, characterized in that, In step 6), the host processing module uses a high dynamic range image display ambient light correction algorithm to correct the high dynamic range image photo synthesized in step 4). Specifically, the high dynamic range image data is tone-mapped to the pixel value range of [0, 255], and then the following adaptive contrast enhancement algorithm is used for correction: 61) Calculate the global luminance mean square error σ for the entire image, and determine the contrast enhancement factor P0 using the following method: If σ≤3, then P0=3. If σ≥10, P0=1, If 3 < σ < 10, then P0 takes linear interpolation, P0 = (27 - 2σ) / 7; 62) Normalize the ambient light intensity to [0,1] to obtain the ambient light intensity factor h. Take P = P0 + K * h, where K is the ambient light intensity amplification parameter. K takes a value between 1 and 3, with 1.5 as the default value. 63) For each pixel I(X,Y), calculate the Gaussian convolution G(X,Y) of that pixel, and calculate its exponent E(X,Y)=(G(X,Y) / I(X,Y)) P ; 64) Calculate the adaptively enhanced pixel value S(X,Y) = 255 * In(X,Y) for each pixel. E(X，Y) , where In(X,Y) is the normalized value of the original image pixel I(X,Y): In(X,Y)=I(X,Y) / 255.

Images