Ghost detection method, device and medium for crystal HUD

Abstract

Embodiments of the present application provide a ghost detection method, device and medium for a crystal HUD. The method comprises: collecting a projection image formed by a crystal HUD projecting a calibration target image; processing the projection image into a target binary image; the foreground of the target binary image comprises a main image and a ghost image in the projection image; performing dynamic step dilation processing on the target binary image using a morphological structure element until the number of connected domains in the image is the same as the preset standard number of independent objects, to obtain a plurality of main image-ghost image composite associations; performing ghost detection on each main image-ghost image composite association to obtain a ghost detection result. The method is used to improve the detection accuracy of HUD ghost detection.

Description

Technical Field

[0001] This application relates to the field of vehicles, and more particularly to a method, device and medium for detecting ghosting in a crystal HUD. Background Technology

[0002] With the evolution of automotive head-up display (HUD) system technology and the application of large-curvature freeform windshields, ghosting detection of HUD projected images has become a core indicator for measuring display quality.

[0003] However, existing projection detection methods are mostly based on the assumption that the primary image (also known as the real image) and the secondary image (also known as the ghost image) are spatially linearly translated and closely connected. When faced with discrete separated double images or non-parallel intersecting ghost images caused by systems with large curvature freeform surfaces, the use of existing methods to perform ghost detection often results in missed or false detections because the primary and secondary images cannot be paired, which seriously affects the detection accuracy of ghost detection.

[0004] Therefore, how to improve the accuracy of HUD ghosting detection under complex optical distortion conditions has become an urgent technical problem to be solved. Summary of the Invention

[0005] This application provides a method, device, and medium for detecting ghosting in a crystal HUD, which solves the problem of missed detection and false detection caused by the inability to match the primary and secondary images under complex distortion conditions in existing methods, thereby improving the detection accuracy of HUD ghosting.

[0006] In a first aspect, embodiments of this application provide a ghosting detection method for a crystal HUD, including:

[0007] Acquire the projected image formed by the crystal HUD projection calibration target;

[0008] The projected image is processed into a target binarized image; the foreground of the target binarized image includes the main image and the secondary image in the projected image;

[0009] The target binarized image is dynamically dilated using morphological structural elements until the number of connected components in the image is the same as the number of preset standard independent objects, resulting in multiple main image-secondary image composite correlation bodies.

[0010] Ghosting detection is performed on each main image-secondary image composite to obtain the ghosting detection results.

[0011] In one possible implementation, processing the projected image into a target binarized image includes:

[0012] The projected image is segmented into a first-order binary image based on a preset threshold.

[0013] Based on the primary binarized image, the target binarized image is obtained.

[0014] In one possible implementation, obtaining the target binarized image based on the primary binarized image includes:

[0015] The overlapping lines in the primary binarized image are decoupled to obtain the decoupled image.

[0016] The decoupled image is determined as the target binarized image.

[0017] In one possible implementation, the process of decoupling the overlapping lines in the primary binarized image to obtain a decoupled image includes:

[0018] Based on the line structure in the calibration target diagram, at least one structural element with directional selectivity is constructed.

[0019] The primary binarized image is subjected to an opening operation using at least one directionally selective structuring element to cut off the bridging regions between overlapping lines in the primary binarized image, thereby obtaining the decoupled image.

[0020] In one possible implementation, before acquiring the target binarized image based on the primary binarized image, the method further includes:

[0021] The primary binarized image is subjected to noise removal processing to eliminate false targets in multiple connected components of the primary binarized image.

[0022] In one possible implementation, the noise removal process on the primary binarized image includes:

[0023] For each connected component in the primary binarized image, calculate the area of ​​the connected component, the aspect ratio of the minimum bounding rectangle, and the skeleton length;

[0024] Pixels in the connected component whose area, aspect ratio of the minimum bounding rectangle, or skeleton length does not meet the preset range requirements of the parameters are set as the background of the primary binarized image.

[0025] In one possible implementation, the ghost detection for each main-secondary image composite to obtain the ghost detection result includes:

[0026] For each main image-secondary image composite, obtain the standard line width reference value corresponding to the main image-secondary image composite;

[0027] Calculate the minimum bounding matrix of the main image-secondary image composite correlation, and extract the width of the short side of the minimum bounding matrix;

[0028] Based on the line width reference value and the short side width, the detection result corresponding to the main image-secondary image composite correlation is obtained. The detection result includes whether there is a ghost in the image region corresponding to the main image-secondary image composite correlation in the ghost image.

[0029] The detection results corresponding to each main image-secondary image composite are summarized to obtain the ghosting detection results.

[0030] In one possible implementation, obtaining a standard line width reference value corresponding to each main-secondary image composite element includes:

[0031] If the line widths in the calibration target image are all the same, then the median of the short side widths corresponding to the multiple main image-sub-image composite correlation bodies is determined as the standard line width reference value.

[0032] In one possible implementation, obtaining the detection result corresponding to the main image-sub-image composite based on the line width reference value and the short side width includes:

[0033] If the width of the short side is greater than the reference value of the line width, then it is determined that there is a ghost in the image region corresponding to the main image-secondary image composite correlation in the ghost image;

[0034] The ghosting index quantization value is calculated based on the difference between the line width reference value and the short side width; the detection result also includes the ghosting index quantization value.

[0035] Secondly, embodiments of this application provide a ghosting detection device for a crystal HUD, comprising:

[0036] The acquisition module is used to acquire the projected image formed by the crystal HUD projection calibration target.

[0037] The first processing module is used to process the projected image into a target binarized image; the foreground of the target binarized image includes the main image and the secondary image in the projected image;

[0038] The second processing module is used to perform dynamic step-size dilation processing on the target binarized image using morphological structural elements until the number of connected components in the image is the same as the number of preset standard independent objects, thereby obtaining multiple main image-secondary image composite correlation bodies.

[0039] The detection module is used to perform ghosting detection on each main image-secondary image composite and obtain the ghosting detection results.

[0040] In one possible implementation, the first processing module includes:

[0041] The processing unit is used to perform first-order binarization segmentation on the projected image according to a preset threshold to obtain a primary binarized image.

[0042] The first acquisition unit is used to acquire the target binarized image based on the primary binarized image.

[0043] In one possible implementation, the acquisition unit is specifically used for:

[0044] The overlapping lines in the primary binarized image are decoupled to obtain the decoupled image.

[0045] The decoupled image is determined as the target binarized image.

[0046] In one possible implementation, the acquisition unit is specifically used for:

[0047] Based on the line structure in the calibration target diagram, at least one structural element with directional selectivity is constructed.

[0048] The primary binarized image is subjected to an opening operation using at least one directionally selective structuring element to cut off the bridging regions between overlapping lines in the primary binarized image, thereby obtaining the decoupled image.

[0049] In one possible implementation, the device further includes:

[0050] The noise reduction module is used to perform noise removal processing on the primary binarized image to remove false targets in multiple connected components of the primary binarized image.

[0051] In one possible implementation, the noise reduction module is specifically used for:

[0052] For each connected component in the primary binarized image, calculate the area of ​​the connected component, the aspect ratio of the minimum bounding rectangle, and the skeleton length;

[0053] Pixels in the connected component whose area, aspect ratio of the minimum bounding rectangle, or skeleton length does not meet the preset range requirements of the parameters are set as the background of the primary binarized image.

[0054] In one possible implementation, the detection module includes:

[0055] The second acquisition unit is used to acquire the standard line width reference value corresponding to each main image-secondary image composite body;

[0056] The calculation unit is used to calculate the minimum bounding matrix of the main image-sub-image composite correlation and extract the short side width of the minimum bounding matrix;

[0057] The third acquisition unit is used to acquire the detection result corresponding to the main image-secondary image composite correlation body according to the line width reference value and the short side width. The detection result includes whether there is a ghost in the image region corresponding to the main image-secondary image composite correlation body in the ghost image.

[0058] The summarization unit is used to summarize the detection results corresponding to each main image-sub-image composite correlation to obtain the ghosting detection results.

[0059] In one possible implementation, the second acquisition unit is specifically used for:

[0060] If the line widths in the calibration target image are all the same, then the median of the short side widths corresponding to the multiple main image-sub-image composite correlation bodies is determined as the standard line width reference value.

[0061] In one possible implementation, the third acquisition unit is specifically used for:

[0062] If the width of the short side is greater than the reference value of the line width, then it is determined that there is a ghost in the image region corresponding to the main image-secondary image composite correlation in the ghost image;

[0063] The ghosting index quantization value is calculated based on the difference between the line width reference value and the short side width; the detection result also includes the ghosting index quantization value.

[0064] Thirdly, embodiments of this application provide a detection and processing device, including: a memory and a processor;

[0065] The memory stores computer-executed instructions;

[0066] The processor executes computer execution instructions stored in the memory, causing the processor to perform the first aspect and / or various possible implementations of the first aspect as described above.

[0067] Fourthly, embodiments of this application provide a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement the first aspect and / or various possible implementations of the first aspect.

[0068] Fifthly, embodiments of this application provide a computer program product, including a computer program that, when executed by a processor, implements the first aspect and / or various possible implementations of the first aspect.

[0069] This application provides a method, device, and medium for ghost detection in a crystal HUD. It acquires a projected image formed by projecting a calibration target image onto the crystal HUD and processes it into a binary image containing both a primary and secondary image in the foreground. This allows for the extraction of ghost detection objects based on a unified image representation. By performing dynamic step-size dilation on the binary image until the number of connected components in the image equals the number of preset standard independent objects, multiple primary-secondary image composite correlations are obtained. This enables stable correlation and effective pairing of primary and secondary images under complex distortion, spatial separation, and non-parallel intersection conditions. Ghost detection is then performed based on each obtained primary-secondary image composite correlation, improving the accuracy of ghost detection. Furthermore, since the dynamic step-size dilation technology used in this method eliminates the need for manual adjustment of algorithm parameters for different vehicle models and windshield surface types, this method effectively ensures the automation level and changeover flexibility of the production line inspection process. Attached Figure Description

[0070] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0071] Figure 1 This is a flowchart illustrating a ghosting detection method for a crystal HUD provided in Embodiment 1 of this application.

[0072] Figure 2 This is a schematic diagram of the separated row and column grid line diagram provided in Embodiment 1 of this application;

[0073] Figure 3 This is a schematic diagram of a projected image provided in Embodiment 1 of this application;

[0074] Figure 4 This is a flowchart illustrating a specific method for ghosting detection in a crystal HUD, as provided in Embodiment 2 of this application.

[0075] Figure 5 This is a flowchart illustrating a specific method for ghosting detection in a crystal HUD as provided in Embodiment 3 of this application.

[0076] Figure 6 This is a schematic diagram of the projected image provided in Embodiment 3 of this application;

[0077] Figure 7 This is a schematic diagram of the structure of a ghosting detection device for a crystal HUD provided in Embodiment 4 of this application;

[0078] Figure 8 This is a schematic diagram of the structure of a ghosting detection device for a crystal HUD provided in Embodiment 5 of this application;

[0079] Figure 9 This is a schematic diagram of the structure of a detection and processing device provided in Embodiment Six of this application.

[0080] The accompanying drawings illustrate specific embodiments of this application, which will be described in more detail below. These drawings and descriptions are not intended to limit the scope of the concept in any way, but rather to illustrate the concept of this application to those skilled in the art through reference to particular embodiments. Detailed Implementation

[0081] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0082] To facilitate understanding of the technical content of this application, the background technology is described in detail below:

[0083] Ghosting detection in Head-Up Display (HUD) systems falls under the field of automotive electronics and image inspection technology, primarily applied in crystal HUD production line inspection scenarios. This scenario typically consists of a HUD projection module, a calibration target image, an imaging acquisition device, and inspection processing equipment. By acquiring the projected image of the calibration target image onto the windshield, the quality of the display status of the primary and secondary images is assessed.

[0084] Existing detection methods are mostly based on the assumption that the primary image (also known as the real image) and the secondary image (also known as the virtual image) are spatially translated linearly and closely connected. Using the standard grid line diagram (i.e., the intersection of horizontal and vertical lines) commonly used in this field as a calibration target diagram, the principle of existing detection methods will be explained:

[0085] Extract the light intensity distribution in the projected image in the horizontal or vertical direction and find the brightest edge of the main and secondary image combination; determine the actual absolute width of the ghosting by the difference between the geometric width of the brighter main image and the total visual width after superimposing the darker secondary image paired with the main image; traverse each group of lines in the detection image and calculate the difference between the widest edge pair and the narrowest edge pair to achieve a quantitative evaluation of the full-screen ghosting value.

[0086] However, if the inspection conditions are complex (such as large curvature windshields, complex optical path distortion, or large assembly deviations), the projected image may show complete spatial separation of the main and secondary images or non-parallel intersecting lines (i.e., cross distortion). In this case, traditional inspection methods often cannot effectively match the main and secondary images, resulting in missed or false detections of line ghosting.

[0087] Therefore, how to achieve stable correlation between the primary and secondary images under complex optical distortion conditions and improve the accuracy of ghost detection results has become an urgent technical problem to be solved.

[0088] To address the aforementioned issues, a ghosting detection method for crystal HUDs is provided. This method involves acquiring a projected image formed by projecting a calibration target image onto the crystal HUD, processing the projected image into a target binarized image with a foreground including a main image and a secondary image, and then performing dynamic step-size dilation on the target binarized image until the number of connected components in the image is equal to the number of preset standard independent objects. This yields multiple main-secondary image composite correlations, and ghosting detection is performed on each composite correlation to improve detection stability and quantization reliability in complex scenes.

[0089] Meanwhile, the ghosting detection method for crystal HUD provided in this application can also achieve the following effects: the method provided in this application does not require manual adjustment of the pairing window or search threshold for different car models and different windshield surface shapes. Through dynamic step expansion technology with a preset number of independent objects as closed-loop feedback control, it can automatically optimize the best algorithm parameters. In this way, it can also achieve the effect of ensuring the automation level of production line inspection and the flexibility of production changeover.

[0090] It should be noted that the execution subject of the ghosting detection method for Crystal HUD provided in this application can be a detection processing device, which can be an industrial control computer, an embedded processor, or a dedicated vision processing module, etc. This application does not limit the specific execution subject of the method.

[0091] The technical solution of this application and how the technical solution of this application solves the above-mentioned technical problems are described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of this application will now be described with reference to the accompanying drawings.

[0092] Figure 1 This is a flowchart illustrating a ghosting detection method for a crystal HUD provided in Embodiment 1 of this application, as shown below. Figure 1 As shown, the method includes:

[0093] S101: Acquires the projected image formed by the crystal HUD projection calibration target.

[0094] Among them, the crystal HUD is the object to be tested. In actual application scenarios, the control unit needs to send a calibration target image display command to the projection module of the crystal HUD in advance, so that the calibration target image is projected onto the windshield through the HUD optical path and forms a projection image.

[0095] In this step, an image acquisition device is needed to capture the projected image formed on the windshield. This image acquisition device can be an industrial camera, an area array image sensor, or other imaging equipment capable of stably acquiring HUD projection results.

[0096] In practical applications, the installation position of the image acquisition device is fixed according to the position calibration result of the HUD, and maintains a predetermined geometric relationship with the windshield and the light output direction of the HUD, so that the acquired image covers the entire display area of ​​the calibration target image.

[0097] In addition, the calibration target image refers to a two-dimensional template image containing a regular calibration pattern; wherein, the calibration target image includes at least one original independent object; when there are multiple original independent objects, each independent object can be a separately set line, frame line, cross mark, rice mark or other target object used for ghosting detection.

[0098] In practical applications, the calibration target plots commonly used in this field are usually standard grid plots, which can be further refined into fully connected grid plots and so on. Figure 2 The diagram shows a split-row and column grid line chart. Specifically, in a fully connected grid line chart, the horizontal and vertical lines intersect completely and are interconnected, forming a continuous grid frame, which is a calibration target chart with a single original independent object; in a split-row and column grid line chart, each line is treated as an original independent object.

[0099] S102: Process the projected image into a target binarized image, wherein the foreground of the target binarized image includes the main image and the secondary image in the projected image.

[0100] In this step, the target binarized image is the result image after foreground extraction from the projected image. Its purpose is to preserve the line regions corresponding to the main image and the secondary image, while suppressing background areas, weak reflection noise, and stray textures from the workstation. Both the main image and the secondary image are retained as foreground components in the target binarized image to ensure that their relative positional relationship can be observed simultaneously during subsequent connected component analysis. In practice, the detection processing equipment can first preprocess the projected image, and then perform foreground extraction and binarization processing on the projected image to obtain the target binarized image.

[0101] It is understandable that, since this application addresses complex optical distortion conditions, the primary image and the secondary image may not necessarily be closely adjacent. Therefore, this step does not force the distinction between the primary and secondary images, but rather preserves them as a whole as a common foreground.

[0102] Based on the above analysis, this step transforms the ghosting detection problem into the identification problem of the foreground connected structure by converting the projected image into a target binarized image. This eliminates the need to rely on the assumption that the main image and the secondary image must be locally continuous or closely distributed.

[0103] S103: Use morphological structural elements to perform dynamic step-size dilation on the target binarized image until the number of connected components in the image is the same as the number of preset standard independent objects, to obtain multiple main image-sub-image composite correlations.

[0104] In this step, morphological structural elements are used to perform dynamic step-size dilation on the foreground region in the target binarized image. After each dilation, the number of connected components in the image is recounted, and the dilation step size for the next round is adjusted according to the relationship between the number of connected components and the number of preset standard independent objects. The dilation step is repeated until the number of connected components is the same as the number of preset standard independent objects, resulting in multiple main image-secondary image composite correlations.

[0105] Among them, morphological structural elements refer to preset shape mask templates used to traverse the foreground region of an image pixel by pixel when performing morphological operations such as dilation. For example, they can be disk-shaped structural elements, circular structural elements, or square structural elements. The step size can be reflected as the radius increment of the structural element, the number of times the dilation is repeated in a single round, or the equivalent dilation scale.

[0106] The main-sub image composite correlation is a detection unit formed by aggregating the main image region and sub image region corresponding to the same standard independent object after dilation. Each main-sub image composite correlation corresponds to a subsequent independent ghosting detection. The number of connected components is the number of independent foreground regions in the image to be dilated. The preset number of standard independent objects is the number of standard independent objects that should exist in the target binarized image. This number is pre-stored according to the target image template before the detection starts. For example, the standard independent objects can be directly the original independent objects in the calibration target image. Correspondingly, the preset number of standard independent objects is directly the number of original independent objects in the calibration target image.

[0107] Taking a disk-shaped structural element as an example, this step may specifically include:

[0108] First, set the initial radius and step gradient of the disk structure element. Then, in each iteration, take the radius and step gradient used in the previous iteration as the radius of the disk structure element in the current iteration, and perform conditional dilation on the image to be dilated to obtain the image after dilation in the current iteration. Then, count the number of connected components in the image. If the number of connected components is equal to the preset number of standard independent objects, stop the iteration. If the number of connected components is greater than the preset number of standard independent objects, proceed to the next iteration.

[0109] The dilation operation is applied to the overall foreground region formed by merging all foreground regions of the main image and all foreground regions of the secondary image. In the first iteration, the image to be dilated is the target binarized image, and the radius of the structuring element used is the initial radius.

[0110] The core principle of this step is that although the main image and secondary image of the same standard independent object may be separated at a distance in the original image, as the dilation scale increases, the graphic regions belonging to the same standard independent object will first bridge to form a joint connected structure. Since the initial spatial positions of different standard independent objects are relatively independent, they usually remain distinct before the target number is reached, thus forming multiple main image-secondary image composite associations that correspond one-to-one with the standard objects.

[0111] For example, Figure 3 This is a schematic diagram of the projected image provided in the embodiments of this application, such as... Figure 3 As shown, the lines in the circled area exhibit complete separation of the primary and secondary images. The method provided in this step bridges the primary and secondary images in the image using a dilation algorithm, achieving effective pairing of primary and secondary images belonging to the same line, thus laying the foundation for improving the detection accuracy of ghosting detection in subsequent steps.

[0112] Based on the above analysis, this step adopts an adaptive morphological iterative association mechanism, which breaks the limitation of spatial continuity and enables the algorithm to accurately associate the primary and secondary images across spatial faults and rotation angles. This effectively solves the problem of the primary and secondary images being completely separated or having axial deflection and intersection, which leads to the inability to pair them and thus results in ghosting, missed detection, or false detection.

[0113] The master-sub image composite correlation obtained by this method provides a stable boundary for subsequent object-by-object ghosting quantization, reducing result fluctuations caused by multi-object confusion. Furthermore, through dynamic compensation expansion processing with preset quantity conditions, while bridging the same pair of master and sub images, it can also, to some extent, prevent spatially adjacent independent standard objects from sticking together. Moreover, this method does not require manual adjustment of the pairing window or search threshold for different vehicle models or windshield surface types, exhibiting strong scene adaptability and significantly improving the automation level and changeover flexibility of production line inspection.

[0114] S104: Perform ghost detection for each main image-secondary image composite to obtain the ghost detection results.

[0115] In this step, ghosting detection is performed line by line on each main image-secondary image composite relational body based on the established object-level association relationship, and the detection results are output separately.

[0116] The ghosting detection results should at least include whether ghosting exists in each main-image-secondary-image composite. Optionally, the ghosting detection results may also include quantitative indicators of the degree of ghosting in each main-image-secondary-image composite with ghosting, such as ghosting width.

[0117] In practical applications, the final ghosting detection results can be directly used as the basis for quality control in the crystal HUD production line. Furthermore, it is understood that since each main-secondary image composite correlation is obtained by converging the number of standard independent objects in S103, the output of this step has a clear object attribution relationship. The method of this application also facilitates the production line's traceability of specific anomaly locations and types.

[0118] This application provides a method, device, and medium for ghost detection in a crystal HUD. It acquires a projected image formed by projecting a calibration target image onto the crystal HUD and processes it into a binary image containing both a primary and secondary image in the foreground. This allows for the extraction of ghost detection objects based on a unified image representation. By performing dynamic step-size dilation on the binary image until the number of connected components in the image equals the number of preset standard independent objects, multiple primary-secondary image composite correlations are obtained. This enables stable correlation and effective pairing of the primary and secondary images under complex distortion, spatial separation, and non-parallel intersection conditions. Ghost detection is then performed based on each obtained primary-secondary image composite correlation, improving the accuracy of ghost detection and enhancing the adaptability and automated detection stability of the crystal HUD to different surface shapes and complex optical conditions. Furthermore, since the dynamic step-size dilation technology used in this method eliminates the need for manual adjustment of algorithm parameters for different vehicle models and windshield surface shapes, this method effectively ensures the automation level and changeover flexibility of the production line inspection process.

[0119] Furthermore, Figure 4 This is a flowchart illustrating a specific ghosting detection method for a crystal HUD provided in Embodiment 2 of this application. Based on the above embodiments, as follows... Figure 4 As shown, this embodiment provides a method, including the following:

[0120] S201: Acquire the projected image formed by the crystal HUD projection calibration target.

[0121] The specific implementation method of this step can be referred to the foregoing embodiments, and will not be repeated here.

[0122] S202: Perform first-order binarization segmentation on the projected image according to the preset threshold to obtain a primary binarized image.

[0123] In this step, based on a preset threshold, the image regions containing the main image and the secondary image in the projected image are set as the foreground in the primary binarized image, and the other image regions in the projected image are set as the background in the primary binarized image.

[0124] For example, each pixel in the projected image can be traversed. If the gray value of the pixel is less than a preset threshold, the gray value of the pixel is set to 0 (i.e., foreground). If the gray value of the pixel is greater than or equal to the preset threshold, the pixel value of the pixel is set to 1 (i.e., background).

[0125] In addition, in practical applications, the preset threshold can be flexibly adjusted according to the local contrast constraints of the projected image, and this application does not impose specific restrictions on it.

[0126] S203: Perform noise removal processing on the primary binarized image to remove false targets in multiple connected components of the primary binarized image.

[0127] In this context, a connected region refers to a connected area belonging to the foreground.

[0128] This step is an optional step in the scheme. Its purpose is to make the foreground structure in the processed binarized image closer to the real line distribution of the calibration target image, reduce the interference of isolated noise points on subsequent processing steps, and thus make the connected component analysis results more stable, thereby improving the stability of the ghost detection process.

[0129] Noise is typically introduced by background stray light, reflections, and lens distortion. Noise removal can specifically include: obtaining the pixel set of each connected component in the binarized image, calculating the geometric features of each connected component, and directly setting connected components that do not meet the preset geometric constraints as background pixels, thereby removing false targets from the primary binarized image.

[0130] In one possible implementation, this step includes the following steps a and b:

[0131] Step a: For each connected component in the primary binarized image, calculate the area of ​​the connected component, the aspect ratio of the minimum bounding rectangle, and the skeleton length.

[0132] In this step, in order to perform boundary constraint-based cascade filtering on each connected component in the primary binarized image, it is necessary to calculate several parameters for each connected component, including area, aspect ratio of the minimum bounding rectangle, and skeleton length.

[0133] For example, for each connected component, the pixel set of the connected component is extracted from the primary binarized image; the total number of pixels in the pixel set is calculated as the area of ​​the connected component; a minimum bounding rectangle is constructed based on the bounding boundary of the connected component, and the ratio of the long side to the short side of the minimum bounding rectangle is calculated to obtain the aspect ratio of the minimum bounding rectangle of the connected component; the connected component is thinned and skeletonized, and the longest path length along the center line of the skeleton is counted as the skeleton length of the connected component.

[0134] Step b: Set the pixels in the connected component whose area, aspect ratio of the minimum bounding rectangle, or skeleton length does not meet the preset range requirements of the parameters as the background of the primary binarized image.

[0135] The preset range requirements for each parameter can be pre-calibrated based on the geometric characteristics of the calibration target image, etc., and this application does not impose any limitations on this.

[0136] As a specific example, only the connected components whose area falls within a preset area range, whose aspect ratio of the minimum bounding matrix falls within a preset aspect ratio range, and whose skeleton hiding degree is greater than or equal to a preset length threshold can be retained as the foreground of the primary binary image, while the remaining connected components are set as the background.

[0137] This implementation uses area, aspect ratio, and skeleton length to jointly constrain connected components, which can simultaneously remove pseudo-targets with abnormal sizes and provide stable input for subsequent processing steps.

[0138] S204: Determine the primary binarized image after denoising as the target binarized image.

[0139] In this step, the primary binarized image is directly identified as the target binarized image to achieve "obtaining the target binarized image based on the secondary binarized image".

[0140] S205: The target binarized image is subjected to dynamic step-size dilation using morphological structural elements until the number of connected components in the image is the same as the number of preset standard independent objects, resulting in multiple main image-sub-image composite correlations, wherein the number of preset standard independent objects is equal to the number of original independent objects in the calibration target image.

[0141] The specific implementation method of this step can be referred to the aforementioned embodiments, and will not be repeated here.

[0142] S206: Perform ghost detection for each main image-secondary image composite to obtain the ghost detection results.

[0143] As a specific example, for each primary-secondary image composite, the standard geometric parameters corresponding to the primary-secondary image composite under ideal imaging conditions without ghosting can be used as a benchmark; by comparing the geometric dimensions of the currently measured composite composite, it can be determined whether ghosting exists. Furthermore, the severity of ghosting can be quantified based on the degree of deviation of the measured parameters from the standard parameters, thereby obtaining the detection result corresponding to the primary-secondary image composite; by summarizing the detection results corresponding to all primary-secondary image composites, the overall ghosting detection result of the crystal HUD projection image is obtained.

[0144] The ghost detection method for crystal HUD provided in this application embodiment can stably preserve the main image and secondary image in the projected image as a binary foreground by performing first-order binarization segmentation on the projected image based on a preset threshold. At the same time, by performing noise reduction processing on the primary binarized image, the interference of background noise on subsequent connected component analysis can be suppressed. Thus, a clear and stable image foundation can be provided for the formation of the subsequent main-secondary image composite correlation and ghost detection.

[0145] Furthermore, Figure 5 This is a flowchart illustrating a specific method for ghosting detection in a crystal HUD as provided in Embodiment 3 of this application, as follows: Figure 5 As shown, the method provided in this application embodiment includes the following:

[0146] S301: Acquires the projected image formed by the crystal HUD projection calibration target.

[0147] S302: Perform first-order binarization segmentation on the projected image according to a preset threshold to obtain a primary binarized image;

[0148] S303: Perform noise removal processing on the primary binarized image to remove false targets in multiple connected components of the primary binarized image.

[0149] The specific implementation methods for steps S301 to S303 can be referred to the foregoing embodiments, and will not be repeated here.

[0150] S304: Decouple the overlapping lines in the denoised primary binarized image to obtain the decoupled image.

[0151] In this step, it is necessary to decouple any two lines in the processed binarized image that have overlapping regions.

[0152] Understandably, decoupling lines in a binarized image can transform overlapping local areas into distinguishable independent lines, maintaining a clearer line distribution in the target binarized image. On one hand, if each original independent object in the calibration target image is a single line, this step can address the phenomenon where adjacent lines (such as intersections of horizontal and vertical lines, or edges of dense lines) merge due to optical diffusion or highlight adhesion. It restores each line to the high-purity state shown in the calibration target image, eliminating line adhesion and ensuring the effectiveness of subsequent dilation processing, further improving detection accuracy. On the other hand, if the original independent objects in the calibration target image are patterns formed by overlapping lines (such as star patterns, crosses, or connected grids), this step effectively removes overlapping interference at intersection points, eliminates highlight halos, and cuts off the coupled lines. This also facilitates subsequent independent detection of individual lines, further improving the precision of ghosting detection.

[0153] In one possible implementation, this step specifically includes steps 4.1 to 4.2 as follows:

[0154] Step 4.1: Based on the line structure in the calibration target image, construct at least one structural element with direction selectivity.

[0155] Step 4.2: Perform an opening operation on the primary binarized image using at least one directionally selective structuring element to cut off the bridging regions between overlapping lines in the primary binarized image, resulting in a decoupled image.

[0156] In practice, the main direction corresponding to the overlapping area is first determined based on the line arrangement of the calibration target image. Then, rectangular, linear, or slender elliptical structural elements are generated according to this main direction and used for erosion and dilation operations on the primary binarized image. Specifically, the erosion operation first removes the narrower connecting parts in the bridging region, while the subsequent dilation operation restores the main area of ​​the lines without reconnecting the severed overlapping parts, thus forming a decoupled image that is separated from the original image.

[0157] As a specific example, for Figure 6 The adhesion of adjacent lines shown in the circled area can be addressed by constructing horizontal and vertical structural elements, respectively, and then performing opening operations (morphological erosion followed by morphological dilation) on the binarized image using these elements to separate the physically connected adjacent line segments. In this example, the opening operations based on horizontal and vertical structural elements can precisely sever the transitional bridging regions where horizontal and vertical lines intersect or are abnormally adhered, while preserving the parallel lines themselves.

[0158] The processing method in this implementation utilizes a directional template consistent with the line structure of the calibration target image to perform directional opening operations on the narrow connecting regions between overlapping lines. This separates the originally connected lines into independent contours in the target binarized image, resulting in a decoupled image that can be used for subsequent ghosting correlation analysis. Therefore, the decoupling result more accurately reflects the true spatial distribution of the main and secondary images, providing stable input for subsequent connected component statistics and ghosting detection, and reducing the interference of overlapping lines on image judgment.

[0159] S305: Determine the decoupled image as the target binarized image.

[0160] In this step, steps S304 and S305 first decouple the primary binarized image by lines, and then determine the decoupled image as the target binarized image, so as to achieve "obtaining the target binarized image based on the secondary binarized image".

[0161] S306: The target binarized image is subjected to dynamic step-size dilation processing using morphological structural elements until the number of connected components in the image is the same as the number of preset standard independent objects, resulting in multiple main image-sub-image composite correlations. The number of preset standard independent objects is equal to the number of independent objects in the calibration target image after line decoupling.

[0162] S307: Perform ghost detection for each main image-secondary image composite to obtain the ghost detection result.

[0163] In one possible implementation, this step may specifically include steps 6.1 to 6.4 as follows:

[0164] Step 6.1: For each main image-secondary image composite, obtain the reference value of the standard line width corresponding to the main image-secondary image composite.

[0165] The standard line reference value refers to the parameter value used to evaluate and quantify whether there is ghosting in the lines of the primary-secondary image composite. For example, it can be the width value of the corresponding line in an ideal imaging state without ghosting.

[0166] As a specific implementation, if the line widths in the calibration target image are all the same, the median of the short side widths corresponding to multiple main-sub-image composite correlation bodies is determined as the standard line width reference value.

[0167] In practice, the minimum bounding rectangle of each main image-secondary image composite body can be calculated, and its short side width can be extracted. Then, the median of the short side widths of all main image-secondary image composite bodies can be calculated, and this median value can be used as the reference value for the standard line width.

[0168] In this embodiment, the line width is determined directly by the median value based on the projected image, eliminating the need to pre-calibrate and store the line width of the ghost-free standard reference image. This simplifies the preparation process before inspection and greatly improves the convenience of on-site inspection.

[0169] Optionally, for a composite association of a primary and secondary image in which the primary and secondary images are completely separated before dynamic step-size expansion, primary and secondary image identification can be performed in advance on each connected region in the binarized image (i.e., the target binarized image) of the primary and secondary image composite association before dynamic step-size expansion, and the width of the minimum outer matrix of the identified primary image connected region can be determined as the standard line reference value.

[0170] For example, the identification of primary and secondary images may include the following:

[0171] For each connected region in the target binarized image, the corresponding pixel set in the projected image is extracted, and the average gray value of the pixel set is calculated. Since the energy of the main image is concentrated and the energy of the secondary image is severely attenuated, the average gray value of the connected regions in the entire image exhibits a typical bimodal distribution. Therefore, the global average threshold of all connected regions in the target binarized image can be used as the decision surface. The global average threshold is obtained by averaging the average gray values ​​of all connected regions in the target binarized image. For each connected region in the target binarized image corresponding to the main-secondary image composite correlation, if the average gray value of the connected region is greater than or equal to the global average threshold, the connected region is determined as the main image region. If the average gray value of the connected region is less than the global average threshold, the connected region is determined as the secondary image region.

[0172] Step 6.2: Calculate the minimum bounding matrix of the primary image-secondary image composite correlation and extract the short side width of the minimum bounding matrix.

[0173] In practice, Green's formula or the Rotating Calipers method can be used to calculate the minimum bounding rectangle of the primary-secondary image composite correlation.

[0174] Step 6.3: Based on the line width reference value and the short side width, obtain the detection result corresponding to the main image-secondary image composite correlation. The detection result includes whether there is a ghost in the image region corresponding to the main image-secondary image composite correlation in the ghost image.

[0175] As a specific implementation method, the detection result also includes a quantified value of the ghosting index. Accordingly, this step includes the following steps 6.3.1 and 6.3.2:

[0176] Step 6.3.1: If the width of the short side is greater than the reference value of the line width, then it is determined that there is a ghost in the image region corresponding to the composite correlation of the main image and the secondary image in the ghost image.

[0177] Step 6.3.2: Calculate the quantified value of the ghosting index based on the difference between the line width reference value and the short side width.

[0178] For example, the ghosting quantization value may include the actual physical width of the ghosting, which is calculated by multiplying the difference between the line width reference value and the short side width with the camera sub-pixel calibration coefficient to determine the actual physical width of the ghosting of the corresponding line in the projected image.

[0179] Optionally, after performing primary and secondary image identification on the connected components corresponding to the primary-secondary image composite correlation in the target binarized image, the angle between the minimum bounding matrix of the primary image and the principal axis normal of the primary-secondary image composite correlation can be determined as the rotational deflection angle caused by the non-parallel ghosting of the primary-secondary image composite correlation; this rotational deflection angle is then used as a quantified value of a ghosting index corresponding to the primary-secondary image composite correlation. In this way, comprehensive quantification of ghosting indexes can be achieved for ghosting situations.

[0180] Understandably, by using the rotation deflection angle as a quantitative value for ghosting indicators, the limitation of traditional algorithms that can only measure one-dimensional translation distance is broken. Through the dual minimum bounding rectangle solution model, not only can the absolute width of the ghosting be output, but also the non-parallel rotation angle and spatial divergence of the main and secondary images can be quantitatively output. This can provide richer and more accurate digital diagnostic basis for the hardware assembly of HUD systems and the improvement of PVB film surface defects in windshields.

[0181] Step 6.4: Summarize the detection results corresponding to each main image-secondary image composite correlation to obtain the ghosting detection results.

[0182] The method provided in steps 6.1 to 6.4 of this implementation directly binds the ghosting judgment logic to the geometric features of the line width of each main image-secondary image composite, and performs independent detection on each group of main image-secondary image composites, which can effectively ensure the refinement of the ghosting detection results and accurately identify the slight ghosting offset defects generated by each line.

[0183] The ghosting detection method for crystal HUDs provided in this application decouples lines in the binarized image, maintaining a clearer line distribution in the target binarized image and transforming originally overlapping local areas into distinguishable independent line knots. This method not only enables individual detection of each line, improving the precision of the detection results, but also effectively eliminates the phenomenon of adjacent lines merging due to optical diffusion or high-brightness adhesion, restoring each line to a high-purity state as shown in the calibration target image, eliminating line adhesion, thereby ensuring the effectiveness of subsequent dilation processing and further improving detection accuracy.

[0184] Figure 7This is a schematic diagram of the structure of a ghosting detection device for a crystal HUD provided in Embodiment 4 of this application, as shown below. Figure 7 As shown, the ghosting detection device 40 for a crystal HUD provided in this embodiment includes:

[0185] Acquisition module 401 is used to acquire the projected image formed by the crystal HUD projection calibration target image;

[0186] The first processing module 402 is used to process the projected image into a target binarized image; the foreground of the target binarized image includes the main image and the secondary image in the projected image;

[0187] The second processing module 403 is used to perform dynamic step-dilation processing on the target binarized image using morphological structural elements until the number of connected components in the image is the same as the number of preset standard independent objects, thereby obtaining multiple main image-sub-image composite correlation bodies.

[0188] The detection module 404 is used to perform ghosting detection on each main image-sub-image composite and obtain the ghosting detection result.

[0189] The ghosting detection device 40 for the crystal HUD provided in this embodiment can execute the method provided in the above method embodiment. Its implementation principle and technical effect are similar, and will not be described in detail here.

[0190] Figure 8 This is a schematic diagram of the ghosting detection device for a crystal HUD provided in Embodiment 5 of this application, as shown below. Figure 8 As shown, based on the above embodiments, the ghosting detection device 40 for a crystal HUD provided in this embodiment further includes:

[0191] The denoising module 405 is used to perform noise removal processing on the primary binarized image to remove false targets in multiple connected components of the primary binarized image.

[0192] In one possible implementation, the first processing module 402 includes:

[0193] The processing unit is used to perform first-order binarization segmentation on the projected image according to a preset threshold to obtain a primary binarized image;

[0194] The first acquisition unit is used to acquire the target binary image based on the primary binary image.

[0195] In one possible implementation, the acquisition unit is specifically used for:

[0196] Line decoupling is performed on the overlapping lines in the primary binarized image to obtain the decoupled image;

[0197] The decoupled image is determined as the target binarized image.

[0198] In one possible implementation, the acquisition unit is specifically used for:

[0199] Based on the line structure in the calibration target diagram, construct at least one structural element with directional selectivity;

[0200] An opening operation is performed on the primary binarized image using at least one directionally selective structuring element to cut off the bridging regions between overlapping lines in the primary binarized image, resulting in a decoupled image.

[0201] In one possible implementation, the noise reduction module 405 is specifically used for:

[0202] For each connected component in the primary binarized image, calculate the area of ​​the connected component, the aspect ratio of the minimum bounding rectangle, and the skeleton length;

[0203] Pixels in a connected component whose area, aspect ratio of the minimum bounding rectangle, or skeleton length does not meet the preset range requirements of the parameters are set as the background of the primary binarized image.

[0204] In one possible implementation, the detection module 404 includes:

[0205] The second acquisition unit is used to acquire the standard line width reference value corresponding to each main image-secondary image composite body;

[0206] The calculation unit is used to calculate the minimum bounding matrix of the main image-secondary image composite correlation and extract the width of the short side of the minimum bounding matrix;

[0207] The third acquisition unit is used to acquire the detection result corresponding to the main image-secondary image composite correlation body based on the line width reference value and the short side width. The detection result includes whether there is a ghost in the image region corresponding to the main image-secondary image composite correlation body in the ghost image.

[0208] The summarization unit is used to summarize the detection results corresponding to each main image-secondary image composite correlation to obtain the ghosting detection results.

[0209] In one possible implementation, the second acquisition unit is specifically used for:

[0210] If all lines in the calibration target image have the same width, the median of the short side widths of multiple main-secondary image composites is determined as the standard line width reference value.

[0211] In one possible implementation, the third acquisition unit is specifically used for:

[0212] If the width of the short side is greater than the reference value of the line width, then it is determined that there is a ghost in the image region corresponding to the composite image of the main image and the secondary image in the ghost image.

[0213] The ghosting index quantification value is calculated based on the difference between the line width reference value and the short side width; the detection result also includes the ghosting index quantification value.

[0214] The ghosting detection device 40 for the crystal HUD provided in this embodiment can execute the method provided in the above method embodiment. Its implementation principle and technical effect are similar, and will not be described in detail here.

[0215] Figure 9 This is a schematic diagram of the structure of a detection and processing device provided in Embodiment Six of this application. Figure 9 As shown, the detection processing device 50 provided in this embodiment includes at least one processor 501 and a memory 502. Optionally, the detection processing device 50 further includes a communication component 503. The processor 501, memory 502, and communication component 503 are connected via a bus 504.

[0216] In a specific implementation, at least one processor 501 executes computer execution instructions stored in memory 502, causing at least one processor 501 to perform the above-described method.

[0217] The specific implementation process of processor 501 can be found in the above method embodiments, and its implementation principle and technical effect are similar. It will not be repeated here.

[0218] In the above embodiments, it should be understood that the processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), etc. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in this invention can be directly implemented by a hardware processor, or implemented by a combination of hardware and software modules within the processor.

[0219] The memory may include read-only memory and random access memory. The memory may be volatile or non-volatile, or may include both. Non-volatile memory may include read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. Volatile memory may include random access memory (RAM), which serves as an external cache. Many forms of RAM are available by way of example, but not limitation. Examples include Static Random Access Memory (SRAM), Dynamic Random Access Memory (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (DDR SDRAM), Enhanced Synchronous DRAM (ESDRAM), Sync Link DRAM (SLDRAM), and Direct Rambus RAM (DR RAM).

[0220] The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of illustration, the buses shown in the accompanying drawings are not limited to a single bus or a single type of bus.

[0221] This application also provides a computer program product, including a computer program that, when executed, implements the above-described method.

[0222] This application also provides a computer-readable storage medium storing computer-executable instructions, which, when executed, implement the above-described method.

[0223] The aforementioned readable storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as SRAM, EEPROM, EPROM, PROM, ROM, magnetic storage, flash memory, magnetic disk, or optical disk. The readable storage medium can be any available medium accessible to a general-purpose or special-purpose computer.

[0224] An exemplary readable storage medium is coupled to a processor, enabling the processor to read information from and write information to the readable storage medium. Of course, the readable storage medium can also be a component of the processor. The processor and the readable storage medium can reside in an ASIC. Alternatively, the processor and the readable storage medium can exist as discrete components in a device.

[0225] The division of units is merely a logical functional division; in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be indirect coupling or communication connection through some interfaces, devices, or units, and may be electrical, mechanical, or other forms.

[0226] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0227] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0228] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, ROM, RAM, magnetic disks, or optical disks.

[0229] Those skilled in the art will understand that all or part of the steps of the above-described method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When executed, the program performs the steps of the above-described method embodiments; and the aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disks, or optical disks.

[0230] The above embodiments are merely preferred embodiments provided to fully illustrate the present invention, and the scope of protection of the present invention is not limited thereto. Equivalent substitutions or modifications made by those skilled in the art based on the present invention are all within the scope of protection of the present invention.

[0231] Finally, it should be noted that other embodiments of the invention will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This invention is intended to cover any variations, uses, or adaptations of the invention that follow the general principles of the invention and include common knowledge or customary techniques in the art not disclosed herein, and is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of the invention is limited only by the appended claims.

Claims

1. A method for detecting ghosting in a crystal HUD, characterized in that, The method includes: Acquire the projected image formed by the crystal HUD projection calibration target; The projected image is processed into a target binarized image; the foreground of the target binarized image includes the main image and the secondary image in the projected image; The target binarized image is dynamically dilated using morphological structural elements until the number of connected components in the image is the same as the number of preset standard independent objects, resulting in multiple main image-secondary image composite correlation bodies. Ghosting detection is performed on each main image-secondary image composite to obtain the ghosting detection results.

2. The method according to claim 1, characterized in that, The step of processing the projected image into a target binarized image includes: The projected image is segmented into a first-order binary image based on a preset threshold. Based on the primary binarized image, the target binarized image is obtained.

3. The method according to claim 2, characterized in that, The step of obtaining the target binarized image based on the primary binarized image includes: The overlapping lines in the primary binarized image are decoupled to obtain the decoupled image. The decoupled image is determined as the target binarized image.

4. The method according to claim 3, characterized in that, The process of decoupling overlapping lines in the primary binarized image to obtain a decoupled image includes: Based on the line structure in the calibration target diagram, at least one structural element with directional selectivity is constructed. The primary binarized image is subjected to an opening operation using at least one directionally selective structuring element to cut off the bridging regions between overlapping lines in the primary binarized image, thereby obtaining the decoupled image.

5. The method according to any one of claims 2 to 4, characterized in that, Before obtaining the target binarized image based on the primary binarized image, the method further includes: The primary binarized image is subjected to noise removal processing to eliminate false targets in multiple connected components of the primary binarized image.

6. The method according to claim 5, characterized in that, The noise removal process for the primary binarized image includes: For each connected component in the primary binarized image, calculate the area of ​​the connected component, the aspect ratio of the minimum bounding rectangle, and the skeleton length; Pixels in the connected component whose area, aspect ratio of the minimum bounding rectangle, or skeleton length does not meet the preset range requirements of the parameters are set as the background of the primary binarized image.

7. The method according to claim 3 or 4, characterized in that, The ghost detection is performed on each main-secondary image composite to obtain the ghost detection results, including: For each main image-secondary image composite, obtain the standard line width reference value corresponding to the main image-secondary image composite; Calculate the minimum bounding matrix of the main image-secondary image composite correlation, and extract the width of the short side of the minimum bounding matrix; Based on the line width reference value and the short side width, the detection result corresponding to the main image-secondary image composite correlation is obtained. The detection result includes whether there is a ghost in the image region corresponding to the main image-secondary image composite correlation in the ghost image. The detection results corresponding to each main image-secondary image composite are summarized to obtain the ghosting detection results.

8. The method according to claim 7, characterized in that, For each primary-secondary image composite, obtaining the standard line width reference value corresponding to the primary-secondary image composite includes: If the line widths in the calibration target image are all the same, then the median of the short side widths corresponding to the multiple main image-sub-image composite correlation bodies is determined as the standard line width reference value.

9. The method according to claim 7, characterized in that, The step of obtaining the detection result corresponding to the main image-secondary image composite based on the line width reference value and the short side width includes: If the width of the short side is greater than the reference value of the line width, then it is determined that there is a ghost in the image region corresponding to the main image-secondary image composite correlation in the ghost image; The ghosting index quantization value is calculated based on the difference between the line width reference value and the short side width; the detection result also includes the ghosting index quantization value.

10. A detection and processing device, characterized in that, include: Memory, processor; The memory stores computer-executed instructions; The processor executes computer execution instructions stored in the memory, causing the processor to perform the method as described in any one of claims 1-9.

11. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement the method as described in any one of claims 1-9.

Images