Test device and test method for detecting stitching defects of a panoramic camera

Abstract

A test device and a test method for detecting stitching defects of a panoramic camera are provided. The test method includes accessing the panoramic camera to obtain a stitched image, wherein the stitched image includes a chart image corresponding to a chart, wherein the chart includes a plurality of black stripes and a plurality of white stripes; generating a defect image marked with stitching defects according to the chart image; and outputting the defect image.

Description

Technical Field

[0001] This invention relates to a testing device and method for detecting stitching defects in panoramic cameras. Background Technology

[0002] Compared to traditional cameras, panoramic cameras can capture panoramic images containing more information. Currently, panoramic imaging is also increasingly being applied to innovative technologies such as virtual reality (VR), smart sightseeing, and driving assistance.

[0003] Panoramic cameras can be equipped with fisheye lenses or multiple lenses. Compared to fisheye lens cameras, multi-lens cameras offer advantages such as a wider angle of view and higher resolution. To generate panoramic images, multi-lens cameras require multiple lenses to acquire multiple images, which are then stitched together into a panoramic image using algorithms. However, issues such as calibration anomalies, housing compression, or tolerance problems may occur during camera assembly, resulting in stitched images containing stitching defects. If these defects cannot be effectively detected before the cameras are shipped, defective products will enter the market. Therefore, developing a method for detecting stitching defects is one of the important issues in this field. Summary of the Invention

[0004] This invention provides a testing device and method for detecting stitching defects in panoramic cameras, which can detect stitching defects in panoramic images.

[0005] This invention discloses a testing apparatus for detecting stitching defects in a panoramic camera, comprising a processor, a transceiver, and a pattern card. The pattern card includes multiple black stripes and multiple white stripes for the panoramic camera to capture. The transceiver is communicatively connected to the panoramic camera. The processor is coupled to the transceiver, wherein the processor is configured to perform: accessing the panoramic camera via the transceiver to acquire a stitched image corresponding to the captured pattern card, wherein the stitched image includes a pattern card image corresponding to the pattern card; generating a defect image marked with stitching defects based on the pattern card image; and outputting the defect image via the transceiver.

[0006] In one embodiment of the present invention, the processor is further configured to perform: filtering the image of the card according to a first grayscale value threshold and a second grayscale value threshold to generate a defective image.

[0007] In one embodiment of the present invention, the first grayscale threshold is greater than the second grayscale threshold, wherein the processor is further configured to perform: filtering a first noise in the image based on the first grayscale threshold, wherein the first grayscale value of the first noise is greater than the first grayscale threshold; and filtering a second noise in the image based on the second grayscale threshold, wherein the second grayscale value of the second noise is less than the second grayscale threshold.

[0008] In one embodiment of the invention, the processor described above is further configured to perform morphological filtering on the image to produce a defective image.

[0009] In one embodiment of the present invention, the morphological filtering described above includes filtering a third noise in the image based on an erosion operation.

[0010] In one embodiment of the present invention, the above-mentioned image includes a striped image, wherein morphological filtering includes: filtering out a fourth noise from the image, wherein the length of the fourth noise in a first direction is greater than a pixel number threshold, wherein the first direction is perpendicular to the striped image.

[0011] In one embodiment of the invention, the processor is further configured to perform: receiving the pixel resolution corresponding to the panoramic camera via a transceiver; and determining a pixel number threshold based on the pixel resolution.

[0012] In one embodiment of the present invention, the pixel resolution is inversely proportional to the pixel number threshold.

[0013] In one embodiment of the invention, the processor is further configured to perform: perform spot detection on the image to detect spots on the image; and remove the spots from the image to produce a defective image.

[0014] In one embodiment of the invention, the processor described above is further configured to perform: deleting spots based on the area of ​​the spots.

[0015] In one embodiment of the invention, the processor described above is further configured to perform: deleting spots according to the ratio of the major and minor axes of the spots.

[0016] In one embodiment of the invention, the processor is further configured to perform: deleting spots in response to a major-minor axis ratio being less than a ratio threshold.

[0017] In one embodiment of the invention, the processor is further configured to perform: obtaining defect parameters based on the defect image; and outputting the defect parameters via a transceiver.

[0018] In one embodiment of the present invention, the above-mentioned defect parameters include at least one of the following: number of defects, area of ​​defects, and location of defects.

[0019] In one embodiment of the present invention, the processor obtains defect parameters based on at least one of the following: an image recognition model and a machine learning model.

[0020] In one embodiment of the invention, the processor is further configured to perform: extracting a map image from the self-stitched image based on the region of interest.

[0021] In one embodiment of the present invention, the plurality of black stripes include a first black stripe and a second black stripe, wherein the plurality of white stripes include a first white stripe, wherein the first white stripe is located between the first black stripe and the second black stripe, and wherein the first white stripe, the first black stripe and the second black stripe are parallel to each other.

[0022] In one embodiment of the invention, the above-described pattern further includes positioning stripes, wherein the positioning stripes are not parallel to the first white stripe.

[0023] In one embodiment of the invention, the testing apparatus further includes a fixing member. The fixing member secures the panoramic camera so that the stitched image includes the map image.

[0024] In one embodiment of the present invention, the above-described testing apparatus further includes a light source. The light source projects light onto the chart.

[0025] The present invention provides a testing method for detecting stitching defects in a panoramic camera, comprising: accessing the panoramic camera to obtain a stitched image corresponding to a map, wherein the stitched image includes a map image corresponding to the map, wherein the map includes a plurality of black stripes and a plurality of white stripes; generating a defect image marked with stitching defects based on the map image; and outputting the defect image.

[0026] Based on the above, the testing device of the present invention can mark stitching defects in the image on the chart for user reference. The testing device can also provide parameters such as the number of defects, the area of ​​the defects, or the location of the defects for user reference. Therefore, technicians on the panoramic camera production line can quickly determine whether the produced panoramic cameras are defective based on the output image of the testing device. Attached Figure Description

[0027] Figure 1 A schematic diagram of a test apparatus for detecting stitching defects in a panoramic camera is shown according to an embodiment of the present invention.

[0028] Figure 2A , 2B 2C illustrates a schematic diagram of the setup of a panoramic camera and a map card according to an embodiment of the present invention.

[0029] Figure 3A A schematic diagram of a drawing card is shown according to an embodiment of the present invention.

[0030] Figure 3B A schematic diagram of a drawing card is shown according to another embodiment of the present invention.

[0031] Figure 4A A schematic diagram illustrating a stitched image without stitching defects is shown according to an embodiment of the present invention.

[0032] Figure 4BAn embodiment of the present invention is illustrated. Figure 4A The grayscale histogram of the image.

[0033] Figure 5A A schematic diagram illustrating a spliced ​​image containing splicing defects is shown according to an embodiment of the present invention.

[0034] Figure 5B An embodiment of the present invention is illustrated. Figure 5A The grayscale histogram of the image.

[0035] Figure 6A A schematic diagram of an enlarged image of a card is shown according to an embodiment of the present invention.

[0036] Figure 6B , 6C Figures 6D and 6E illustrate a filtered image of a card according to an embodiment of the present invention.

[0037] Figure 7 A flowchart illustrating a test method for detecting stitching defects in a panoramic camera is shown in one embodiment of the present invention.

[0038] Explanation of reference numerals in the attached figures

[0039] 100: Testing equipment;

[0040] 110: Processor;

[0041] 120: Storage medium;

[0042] 130: Transceiver;

[0043] 140: Fastener;

[0044] 150, 1501, 1502: Picture cards;

[0045] 151, 153: White stripes;

[0046] 152, 154: Black stripes;

[0047] 158: Positioning stripes;

[0048] 160: Light source;

[0049] 200: Panoramic camera;

[0050] 210, 220, 230, 240: Lenses;

[0051] 400, 500: stitched images;

[0052] 410, 420, 510, 520: Areas;

[0053] 43, 53, 531: Regions of interest;

[0054] 430, 530, 630, 730, 830, 930: Tuka images;

[0055] 450, 550: Grayscale histogram;

[0056] 451, 452, 453, 551, 552, 553: grayscale value areas;

[0057] 631, 731, 831, 931: Noise;

[0058] D1, D2: Direction;

[0059] S701, S703, S705: Steps;

[0060] T1, T2: Gray level thresholds. Detailed Implementation

[0061] Reference will now be made in detail to exemplary embodiments of the invention, examples of which are illustrated in the accompanying drawings. Wherever possible, the same element symbols are used in the drawings and description to denote the same or similar parts.

[0062] Figure 1 A schematic diagram of a testing apparatus 100 for detecting stitching defects in a panoramic camera is illustrated according to an embodiment of the present invention, wherein the panoramic camera may include a wide-angle multi-lens camera. The testing apparatus 100 may include a processor 110, a storage medium 120, a transceiver 130, a fixture 140, and a plotter 150. In one embodiment, the testing apparatus 100 may further include a light source 160.

[0063] Processor 110 may be, for example, a central processing unit (CPU), or other programmable general-purpose or special-purpose microcontroller (MCU), microprocessor, digital signal processor (DSP), programmable controller, application-specific integrated circuit (ASIC), graphics processing unit (GPU), image signal processor (ISP), image processing unit (IPU), arithmetic logic unit (ALU), complex programmable logic device (CPLD), field programmable gate array (FPGA), or other similar components or combinations thereof. Processor 110 may be coupled to storage medium 120 and transceiver 130, and access and execute multiple modules and various applications stored in storage medium 120.

[0064] Storage medium 120 may be any type of fixed or removable random access memory (RAM), read-only memory (ROM), flash memory, hard disk drive (HDD), solid state drive (SSD), or similar components or combinations thereof, for storing multiple modules or various applications that can be executed by processor 110.

[0065] Transceiver 130 transmits and receives signals wirelessly or via a wired connection. Transceiver 130 can also perform operations such as low-noise amplification, impedance matching, mixing, up or down frequency conversion, filtering, amplification, and similar functions.

[0066] The fastener 140 can be used to secure a panoramic camera or a map card 150 so that the stitched image captured by the panoramic camera includes a map card image corresponding to the map card 150. The map card 150 can emit light when the panoramic image is captured. In one embodiment, the map card 150 may include a backlight panel. The map card 150 can emit light itself through the backlight panel. In one embodiment, a light source 160 can project light onto the map card 150 to cause the map card 150 to emit light.

[0067] Figure 2A , 2B A schematic diagram illustrating the arrangement of the panoramic camera 200 and the map card 150 according to an embodiment of the present invention is shown in Figure 2C. (Refer to...) Figure 2A In this embodiment, the panoramic camera 200 may include lenses 210 and 220. The panoramic camera 200 and the image card 150 may be disposed on the surface of the fixing member 140 such that the image card 150 is located at the overlap of the viewing angles of lenses 210 and 220. Thus, when the panoramic camera 200 captures the image card 150 through lenses 210 and 220 and generates a stitched image, the stitching area in the stitched image, i.e., the region of interest (ROI), may include the image card corresponding to the image card 150. The testing device 100 can extract the image card corresponding to the image card 150 from the stitched image based on the ROI. The testing device 100 can detect whether there are stitching defects in the stitching area corresponding to lenses 210 and 220 based on the image card corresponding to the image card 150.

[0068] Reference Figure 2B In this embodiment, the panoramic camera 200 may include lenses 210, 220, and 230, and the image card 150 may include image card 1501 and image card 1502. The panoramic camera 200, image card 1501, and image card 1502 may be disposed on the surface of the fixing member 140, such that image card 1501 is located at the overlap of the viewing angles of lens 210 and lens 220, and image card 1502 is located at the overlap of the viewing angles of lens 220 and lens 230. Thus, when the panoramic camera 200 captures a stitched image through lenses 210, 220, and 230, a region of interest in the stitched image may include an image card corresponding to image card 1501, and another region of interest in the stitched image may include an image card corresponding to image card 1502. The testing device 100 can detect whether there are stitching defects in the stitched image corresponding to lens 210 and lens 220 based on the image corresponding to image card 1501, and can detect whether there are stitching defects in the stitched image corresponding to lens 220 and lens 230 based on the image corresponding to image card 1502.

[0069] Reference Figure 2CIn this embodiment, the panoramic camera 200 may include lenses 210, 220, 230, and 240, and the image card 150 may include image card 1501 and image card 1502. The panoramic camera 200, image card 1501, and image card 1502 may be disposed on the surface of the fixing member 140, such that image card 1501 is located at the overlap of the viewing angles of lens 210 and lens 220, and image card 1502 is located at the overlap of the viewing angles of lens 220 and lens 230. Thus, when the panoramic camera 200 captures a stitched image through lenses 210, 220, and 230, a region of interest in the stitched image may include an image card image corresponding to image card 1501, and another region of interest in the stitched image may include an image card image corresponding to image card 1502. The testing device 100 can detect whether there are stitching defects in the stitched image corresponding to lens 210 and lens 220 based on the image corresponding to image card 1501, and can detect whether there are stitching defects in the stitched image corresponding to lens 220 and lens 230 based on the image corresponding to image card 1502.

[0070] Next, the user can reset the panoramic camera 200 (e.g., rotate the panoramic camera 200 clockwise by 180 degrees) so that map card 1501 is located at the overlap of the viewpoints of lens 230 and lens 240, and map card 1502 is located at the overlap of the viewpoints of lens 240 and lens 210. Thus, when the panoramic camera 200 captures a stitched image through lenses 210, 230, and 240, one region of interest in the stitched image may include the map card image corresponding to map card 1501, and another region of interest in the stitched image may include the map card image corresponding to map card 1502. The testing device 100 can detect whether there are stitching defects in the stitched image corresponding to lenses 230 and 240 based on the map card image corresponding to map card 1501, and can detect whether there are stitching defects in the stitched image corresponding to lenses 240 and 210 based on the map card image corresponding to map card 1502. In this way, users can complete the testing process of the panoramic camera 200 with multiple lenses in a shorter time.

[0071] Please also refer to Figure 1 and Figures 2A to 2C The transceiver 130 of the test apparatus 100 is communicatively connected to the panoramic camera 200. The processor 110 can access the panoramic camera 200 through the transceiver 130 to obtain a stitched image generated by the panoramic camera 200 capturing the map 150. The stitched image may contain a map image corresponding to the map 150.

[0072] Figure 3AA schematic diagram of a card 150 is illustrated according to an embodiment of the present invention. Card 150 may include a plurality of black stripes and a plurality of white stripes. For example, card 150 may include a white stripe 151, a black stripe 152, a white stripe 153, and a black stripe 154. Black stripe 152 may be located between white stripe 151 and white stripe 153. White stripe 151, black stripe 152, and white stripe 153 may be parallel to each other. White stripe 153 may be located between black stripe 152 and black stripe 154. Black stripe 152, white stripe 153, and black stripe 154 may be parallel to each other.

[0073] Figure 3B A schematic diagram of a pattern card 150 is illustrated according to another embodiment of the present invention. In one embodiment, the pattern card 150 may further include positioning stripes 158. The positioning stripes 158 may not be parallel to the white stripe 151, black stripe 152, white stripe 153, or black stripe 154. The processor 110 may position the pattern card 150 according to the positioning stripes 158. For example, inclined positioning stripes 158 may be interspersed between multiple horizontal stripes (e.g., black or white stripes) of the pattern card 150. The processor 110 may determine whether the splicing of the pattern card 150 is normal according to the positioning stripes 158, avoiding mistaking a splicing defect where "the vertical misalignment is exactly one stripe spacing" as normal.

[0074] Figure 4A A schematic diagram of a stitched image 400 without stitching defects is illustrated according to an embodiment of the present invention. After acquiring two images through lenses 210 and 220 respectively, processor 110 can perform alpha blending on the two images to stitch them together and generate stitched image 400. Specifically, the two images acquired by lenses 210 and 220 may partially overlap. To make the overlap smooth, processor 110 can first make the two images transparent, and then blend the two images according to equation (1) to generate stitched image 400, where RGB1 can represent pixels in the overlap corresponding to the image acquired by lens 210, RGB2 can represent pixels in the overlap corresponding to the image acquired by lens 220, and RGB3 can represent pixels in stitched image 400. Stitched image 400 may include a region 410 corresponding to the viewpoint of lens 210 and a region 420 corresponding to the viewpoint of lens 220. Region of interest 43 may be located at the overlap of regions 410 and 420 (i.e., the stitching area). Region of interest 43 may include the map image 430 corresponding to map 150. If map image 430 does not contain stitching defects, the grayscale value of map image 430 may be as follows: Figure 4B As shown.

[0075] RGB3=(1-a)*RGB1+a*RGB2…(1)

[0076] Figure 4B An embodiment of the present invention is illustrated. Figure 4A The grayscale histogram 450 of the image 430. The grayscale histogram 450 may include a grayscale value region 451 representing white pixels, a grayscale value region 452 representing gray pixels, and a grayscale value region 453 representing black pixels. Figure 4B As can be seen, since image 430 does not contain stitching defects, the black and white stripes will not be mixed together when processor 110 performs alpha synthesis on the two images. Therefore, the number of pixels in grayscale region 452 is very small.

[0077] Figure 5A A schematic diagram of a stitched image 500 containing stitching defects is illustrated according to an embodiment of the present invention. After acquiring two images through lenses 210 and 220 respectively, processor 110 can perform alpha synthesis on the two images to stitch them together and generate stitched image 500. Stitched image 500 may include a region 510 corresponding to the viewpoint of lens 210 and a region 520 corresponding to the viewpoint of lens 220. Region of interest 53 may be located at the overlap of region 510 and region 520 (i.e., the stitching region). Region of interest 53 may include a map image 530 corresponding to map 150. If map image 530 contains stitching defects, the grayscale value of map image 530 may be as follows: Figure 5B As shown.

[0078] Figure 5B An embodiment of the present invention is illustrated. Figure 5A The grayscale histogram 550 of the image 530. The grayscale histogram 550 may include a grayscale value region 551 representing white pixels, a grayscale value region 552 representing gray pixels, and a grayscale value region 553 representing black pixels. Figure 5B As can be seen, because image 530 contains stitching defects, when processor 110 performs alpha synthesis on the two images, the black and white stripes will be mixed together. Therefore, compared to... Figure 4B The grayscale value areas 452 and 552 shown have a very large number of pixels.

[0079] Figure 6AA schematic diagram of an enlarged image 530 is illustrated according to an embodiment of the present invention. The image 530 may include black stripes and white stripes. A splicing defect in the image 530 is located in a selected area 531. Compared to other areas of the image 530, which are mostly composed of black or white pixels, area 531 contains more gray pixels. To make the splicing defect more noticeable, the processor 110 can filter noise in the image 530 based on grayscale thresholds T1 and T2, thereby producing a result such as... Figure 6B Image 630 shown.

[0080] Specifically, the storage medium 120 can pre-store grayscale thresholds T1 and T2, where grayscale threshold T1 can be greater than grayscale threshold T2. Grayscale thresholds T1 and T2 are related to light source brightness and environmental conditions such as the image card. Under fixed environmental conditions, the processor 110 can filter noise in the image card image 530 based on grayscale threshold T1, where the grayscale value of the noise is greater than grayscale threshold T1. For example, noise with a grayscale value greater than grayscale threshold T1 may include white stripes that do not contain splicing defects. On the other hand, the processor 110 can filter noise in the image card image 530 based on grayscale threshold T2, where the grayscale value of the noise is less than grayscale threshold T2. For example, noise with a grayscale value less than grayscale threshold T2 may include black stripes that do not contain splicing defects. Thus, the processor 110 can filter out black or white stripes in the image card image 530 based on grayscale thresholds T1 and T2.

[0081] Figure 6B , 6C Figures 6D and 6E illustrate a filtered card image according to an embodiment of the present invention. After the processor 110 filters noise from the card image 530 according to grayscale thresholds T1 and T2 to generate a card image 630, the card image 630 may include stripe images (e.g., such as...) in the card image 630. Figure 3A The white stripes 151 or black stripes 152 shown are approximately parallel to the direction D1 of the noise 631.

[0082] Processor 110 can perform morphological filtering on image 630 to produce filtered image 730. Specifically, processor 110 can filter out noise 631 from image 630 in response to noise 631 in direction D2 having a length greater than a pixel count threshold, where direction D2 is perpendicular to direction D1. For example, the pixel count threshold could be 2. If the length of noise 631 in direction D2 is greater than 2 pixels, processor 110 can consider noise 631 as striped noise. Accordingly, processor 110 can filter out noise 631 from image 630.

[0083] One or more pixel count thresholds may be pre-stored in storage medium 120. In one embodiment, the pixel count threshold may be associated with the pixel resolution of panoramic camera 200. Processor 110 may receive the pixel resolution corresponding to panoramic camera 200 via transceiver 130 and determine the pixel count threshold based on the pixel resolution. When two images are stitched or alpha-combined, the stitched image produced by the image with higher resolution will have narrower stripe noise; therefore, the pixel count threshold may be inversely proportional to the pixel resolution. Processor 110 may select a pixel count threshold suitable for panoramic camera 200 from storage medium 120 based on pixel resolution.

[0084] like Figure 6C As shown, image 730 may contain small areas of noise 731. Noise 731 may be referred to as salt and pepper noise. Processor 110 may perform morphological filtering on image 730 to produce a filtered image 830. Processor 110 may perform an open operation on image 730 to filter out noise 731 in image 730. Specifically, processor 110 may first perform an erosion operation on image 730, and then perform a dilation operation on image 730, thereby filtering out noise 731 in image 730 to produce image 830. The open operation can smooth the contours of objects in image 730 and filter out salt and pepper noise and narrow fine lines in image 730. In other words, the open operation can break the small links between objects, making two objects clearly distinguishable.

[0085] like Figure 6D As shown, the image 830 may contain blob-like noise 831. The processor 110 may perform blob detection on the image 830 to detect blobs (i.e., noise 831) on the image 830. The processor 110 may then remove the blobs from the image 830, resulting in a filtered image 930. In one embodiment, the processor 110 may determine whether to remove a blob based on the ratio of its major and minor axes. For example, suppose the noise 831 is an elliptical blob. The ratio of its major and minor axes is the ratio of the major axis to the minor axis of the ellipse. The processor 110 may remove the noise 831 in response to the ratio of its major and minor axes being less than a ratio threshold. The ratio threshold may be pre-stored in the storage medium 120.

[0086] like Figure 6EAs shown, image 930 may contain speckled noise 931. Processor 110 may perform speckle detection on image 930 (or image 830) to detect speckles (i.e., noise 931) on image 930. Processor 110 may then remove the speckles from image 930. In one embodiment, processor 110 may determine whether to remove a speckle based on its area. For example, processor 110 may remove noise 931 from image 930 in response to noise 931 having an area greater than an area threshold. The area threshold may be pre-stored in storage medium 120. In one embodiment, the area threshold may be associated with the height of horizontal stripes on image 150. After the erosion operation is completed, since defects higher than the horizontal stripe height resulting from alpha blending have been filtered out, the height of any remaining defects will not exceed the height of the horizontal stripes. Therefore, processor 110 may set the area threshold based on the height of the horizontal stripes of image 150, wherein the area threshold may be proportional to the height of the horizontal stripes. In one embodiment, the area threshold may be associated with the width of the stitching area formed by alpha blending, wherein the area threshold may be proportional to the width of the stitching area. Since the width of the stitching area is an internal parameter of the panoramic camera 200, the processor 110 may pre-generate the area threshold according to the specifications of the panoramic camera 200 and store the area threshold in the storage medium 120.

[0087] It is worth noting that this invention is not limiting. Figures 6A to 6E The relevant paragraphs reveal the execution order of the filtering process. For example, processor 110 may first perform the following on the image card: Figure 6C The erosion operation disclosed in the embodiments is then performed on the map image as follows: Figure 6B The embodiments disclosed herein employ a pixel-number threshold-based filtering process. After performing the various filtering processes described above, noise on the image card can be filtered out. Therefore, the graphic displayed on the filtered image card is the location of the stitching defect. The processor 110 can generate a defect image marked with stitching defects based on the filtered image card and can output the defect image via the transceiver 130 for user reference. The user can quickly determine the location of the stitching defect in the panoramic camera 200 based on the defect image.

[0088] The processor 110 can further obtain defect parameters from the defect image and output the defect parameters through the transceiver 130 for user reference. The defect parameters may include information such as the number of defects, defect area, or defect location, but the invention is not limited thereto. For example, the processor 110 can input the defect image into an image recognition model or a machine learning model to generate defect parameters.

[0089] Figure 7 A flowchart illustrating a test method for detecting stitching defects in a panoramic camera, according to an embodiment of the present invention, wherein the test method may be performed by, for example... Figure 1 The test apparatus 100 shown is implemented. In step S701, a panoramic camera is accessed to obtain a stitched image, wherein the stitched image includes a map image corresponding to a map, wherein the map includes multiple black stripes and multiple white stripes. In step S703, a defect image marked with stitching defects is generated based on the map image. In step S705, the defect image is output.

[0090] In summary, the testing apparatus of the present invention provides a special image card for the panoramic camera under test to capture images. The testing apparatus can filter various noises in the image card image through methods such as grayscale threshold configuration, morphological filtering, pixel number threshold configuration, speckle detection, or scale threshold configuration. If noise still exists in the filtered image card image, it indicates that the image card image contains stitching defects. The testing apparatus can mark the stitching defects in the image card image for user reference. Furthermore, the testing apparatus can also provide parameters such as the number of defects, defect area, or defect location for user reference.

[0091] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A testing device for detecting stitching defects in panoramic cameras, characterized in that, include: The map includes multiple black stripes and multiple white stripes for the panoramic camera to capture; Transceiver, communicatively connected to the panoramic camera; as well as A processor, coupled to the transceiver, wherein the processor is configured to perform: The transceiver accesses the panoramic camera to obtain a stitched image corresponding to the captured map card, wherein the stitched image includes a map card image corresponding to the map card; Generating a defect image marked with the splicing defect based on the image card, including: filtering a first noise in the image card based on a first grayscale value threshold, wherein the first grayscale value of the first noise is greater than the first grayscale value threshold; and filtering a second noise in the image card based on a second grayscale value threshold, wherein the second grayscale value of the second noise is less than the second grayscale value threshold, wherein the first grayscale value threshold is greater than the second grayscale value threshold, wherein the first noise and the second noise do not include the splicing defect; and The defective image is output through the transceiver.

2. The testing apparatus of claim 1, wherein the processor is further configured to perform: Morphological filtering is performed on the image to produce the defective image.

3. The testing apparatus according to claim 2, wherein the morphological filtering comprises: The third noise in the image is filtered out based on the erosion operation.

4. The testing apparatus according to claim 2, wherein the image on the card includes a stripe image, and wherein the morphological filtering includes: A fourth noise is filtered out from the image, wherein the length of the fourth noise in a first direction is greater than a pixel number threshold, wherein the first direction is perpendicular to the stripe image.

5. The testing apparatus of claim 4, wherein the processor is further configured to perform: The transceiver receives the pixel resolution corresponding to the panoramic camera; and The pixel count threshold is determined based on the pixel resolution.

6. The testing apparatus according to claim 5, wherein the pixel resolution is inversely proportional to the pixel number threshold.

7. The testing apparatus of claim 1, wherein the processor is further configured to perform: Perform spot detection on the image card to detect spots on the image card; and The spots are removed from the image to produce the defect image.

8. The testing apparatus of claim 7, wherein the processor is further configured to perform: The spot is deleted based on its area.

9. The testing apparatus of claim 7, wherein the processor is further configured to perform: The spots are deleted according to the ratio of their major and minor axes.

10. The testing apparatus of claim 9, wherein the processor is further configured to perform: The spot is deleted in response to the ratio of the major and minor axes being less than a ratio threshold.

11. The testing apparatus of claim 1, wherein the processor is further configured to perform: Defect parameters are obtained from the defect image; and The defect parameters are output through the transceiver.

12. The testing apparatus of claim 11, wherein the defect parameter includes at least one of the following: The number of defects, the area of ​​defects, and the location of defects.

13. The testing apparatus of claim 11, wherein the processor obtains the defect parameter according to at least one of the following: Image recognition models and machine learning models.

14. The testing apparatus of claim 1, wherein the processor is further configured to perform: The image card is extracted from the stitched image based on the region of interest.

15. The testing apparatus according to claim 1, wherein the plurality of black stripes includes a first black stripe and a second black stripe, wherein the plurality of white stripes includes a first white stripe, wherein the first white stripe is located between the first black stripe and the second black stripe, and wherein the first white stripe, the first black stripe and the second black stripe are parallel to each other.

16. The testing apparatus of claim 15, wherein the chart further includes positioning stripes, wherein the positioning stripes are not parallel to the first white stripe.

17. The testing apparatus according to claim 1, further comprising: A fixing element secures the panoramic camera so that the stitched image includes the map image.

18. The testing apparatus according to claim 1, further comprising: A light source projects light onto the card.

19. A test method for detecting stitching defects in panoramic cameras, characterized in that, include: Access the panoramic camera to obtain a stitched image corresponding to the map card, wherein the stitched image includes a map card image corresponding to the map card, wherein the map card includes multiple black stripes and multiple white stripes; Generating a defect image marked with the splicing defect based on the image card, including: filtering a first noise in the image card based on a first grayscale value threshold, wherein the first grayscale value of the first noise is greater than the first grayscale value threshold; and filtering a second noise in the image card based on a second grayscale value threshold, wherein the second grayscale value of the second noise is less than the second grayscale value threshold, wherein the first grayscale value threshold is greater than the second grayscale value threshold, wherein the first noise and the second noise do not include the splicing defect; and Output the image of the defect.

Images