Color balancing method, system, medium and equipment for vehicle panoramic look-around image

Abstract

The invention discloses a color equalization method, system, medium and equipment for a vehicle panoramic look-around image, and the method comprises the steps: processing images shot by adjacent cameras of a vehicle, obtaining a reference image and a target image, extracting an overlapping region of the reference image and the target image, and obtaining a reference overlapping region image and a target overlapping region image; respectively segmenting the reference coincidence area map and the target coincidence area map to obtain a plurality of first coincidence maps and second coincidence maps; calculating similarity values of the first coincidence graphs and the corresponding second coincidence graphs, and taking the first coincidence graphs and the corresponding second coincidence graphs of which the similarity values accord with a preset threshold range as a first determined coincidence graph and a second determined coincidence graph; and processing the target image according to the color features of the first determined coincidence image and the second determined coincidence image to complete color equalization processing. According to the invention, through a small module segmentation point selection and similarity judgment technology, the problem of color inconsistency of the panoramic image of the vehicle panoramic look-around system under different illumination conditions and complex scenes is solved.

Description

Technical Field

[0001] The present invention belongs to the technical field of color balancing of a vehicle panoramic surround view system, and in particular relates to a color balancing method, system, medium and device for a vehicle panoramic surround view image. Background Art

[0002] With the rapid development of the automotive industry, driver assistance systems play an increasingly important role in improving driving safety and user experience. Among them, the vehicle's panoramic surround view system has attracted widespread attention because it can provide drivers with an all-round view without blind spots. However, in practical applications, due to changes in external lighting conditions (such as day and night, cloudy and sunny days, etc.) and the inherent differences between different cameras, there is a problem of color consistency between the captured images, which directly affects the overall performance of the system and user experience.

[0003] To solve this problem, researchers have developed a series of color balancing algorithms for the vehicle's 360-degree panoramic view system. These algorithms include but are not limited to adaptive exposure control, white balance processing, and color correction, which improve the authenticity and consistency of image colors by intelligently analyzing and adjusting image data from all directions. However, there are still some technical challenges in existing technical means, such as image processing under complex lighting conditions or in complex scenes, and how to maintain high color fidelity while ensuring good visual effects. Summary of the invention

[0004] In response to the shortcomings of the prior art, the present application provides a color balancing method, system, medium and device for vehicle panoramic surround view images. Through small module segmentation point selection and similarity judgment technology, the problem of color inconsistency of panoramic images of vehicle panoramic surround view systems under different lighting conditions and complex scenes is solved.

[0005] To achieve the above objectives, in a first aspect, the present invention provides a color balancing method for a vehicle panoramic surround view image, comprising:

[0006] Preprocessing the original images taken by the adjacent cameras of the vehicle to obtain a reference image and a target image, and extracting the overlapping area of ​​the reference image and the target image to obtain a reference overlapping area image and a target overlapping area image;

[0007] The reference overlap area map is segmented to obtain a plurality of first overlap areas; the target overlap area map is segmented to obtain a plurality of second overlap areas; wherein each of the first overlap areas is overlapped with each of the second overlap areas. Figure 1 One to one correspondence;

[0008] According to a preset algorithm, respectively, a similarity value is calculated for each of the first coincidence graphs and the corresponding second coincidence graphs until the similarity value meets a preset threshold range, and the first coincidence graph and the corresponding second coincidence graph whose similarity value meets the preset threshold range are used as the first determined coincidence graph and the second determined coincidence graph;

[0009] The target image is processed according to the color features of the first determined coincidence image and the second determined coincidence image to complete the color balancing process.

[0010] Compared with the prior art, the embodiments of the present application have the following beneficial effects: pre-processing the images taken by adjacent cameras to obtain a reference image and a target image, and extracting the overlapping area between them, thereby ensuring that subsequent processing is based on the content under the same perspective, which helps to improve the accuracy of color balancing, and by focusing on the overlapping area, the color differences between different cameras can be more accurately identified and adjusted; the reference overlapping area image and the target overlapping area image are respectively divided into multiple small area images, and the local color differences can be captured more finely through fine-grained segmentation, making the subsequent color correction more refined and accurate, and improving the sensitivity to subtle changes in complex scenes; by evaluating the similarity between the segmented images, the actual effective area that can be used for color balancing can be quickly screened out; by processing and adjusting the color features, the accuracy and consistency of the color balancing process are guaranteed.

[0011] In some embodiments of the first aspect of the present application, the similarity value calculation is performed on each of the first coincidence graphs and the corresponding second coincidence graphs according to a preset algorithm, including:

[0012] Calculating image hash values ​​of the first coincidence image and the corresponding second coincidence image respectively;

[0013] The Hamming distance of the hash value of the image is calculated as a similarity value between the corresponding first coincidence image and the second coincidence image.

[0014] The above embodiment has the following beneficial effects: by calculating the image hash value and the Hamming distance of the first coincidence image and the corresponding second coincidence image as the similarity value, it is possible to effectively determine whether the two coincidence images are from the same plane or scene, thereby ensuring the accuracy of subsequent color adjustment, avoiding the color distortion problem caused by non-coplanarity, and enhancing the robustness and adaptability of the algorithm.

[0015] In some embodiments of the first aspect of the present application, processing the target image according to the color features of the first determined coincidence image and the second determined coincidence image to complete the color balancing process includes:

[0016] Respectively obtain the average values ​​of the first determined coincidence image and the second determined coincidence image on each channel of the RGB three channels;

[0017] Calculate the difference between the average values ​​of the first determined coincidence graph and the second determined coincidence graph on each channel according to the average values ​​on each channel;

[0018] The target image is processed according to the difference between the average values ​​on each channel to complete the color balancing process.

[0019] The above embodiment has the following beneficial effects: by obtaining and determining the average value of the overlapped image on the three RGB channels, calculating the difference and processing the target image accordingly, the image color can be accurately adjusted, so that the spliced ​​image is visually more natural and harmonious, and the color consistency of the image is improved.

[0020] In some embodiments of the first aspect of the present application, the preprocessing of the original images captured by the cameras adjacent to the vehicle to obtain the reference image and the target image includes:

[0021] Obtain the original image taken by the vehicle's adjacent cameras;

[0022] Processing each of the original images according to a preset calibration algorithm of the surround view system to obtain the internal and external parameters and distortion coefficients of the camera;

[0023] Processing each of the original images according to a preset dedistortion algorithm of the surround view system and the distortion coefficient to obtain a dedistorted image;

[0024] According to the preset perspective transformation algorithm of the surround view system and the internal and external parameters, each of the dedistorted images is processed to obtain each top view corresponding to the dedistorted image, which is used as a reference image and a target image respectively.

[0025] The above embodiment has the following beneficial effects: the original image is processed by basic steps such as calibration, dedistortion and perspective transformation, which ensures that the images captured by each camera can be accurately spliced ​​in the same coordinate system, laying the foundation for subsequent color balancing processing, and effectively solving the image differences caused by inconsistent camera installation heights, thereby improving the quality and reliability of image processing.

[0026] In some embodiments of the first aspect of the present application, extracting the overlapping area of ​​the reference image and the target image to obtain a reference overlapping area map and a target overlapping area map includes:

[0027] splicing the reference image and the target image according to a preset stitching algorithm of the surround view system to obtain a fused image;

[0028] The overlapping area of ​​the fused image is intercepted, and the portion of the reference image corresponding to the overlapping area is used as a reference overlapping area map, and the portion of the target image corresponding to the overlapping area is used as a target overlapping area map.

[0029] The above embodiment has the following beneficial effects: by splicing the reference image and the target image to obtain a fused image and cutting out the overlapping area therefrom, it is ensured that subsequent processing is performed on the actual overlapping part, which helps to improve the color balancing effect.

[0030] In a second aspect, an embodiment of the present application further provides a color balancing system for a vehicle panoramic surround view image, comprising: an overlap region acquisition module, a segmentation module, an overlap map determination module, and a processing module;

[0031] The overlap region acquisition module is used to pre-process the original images taken by the adjacent cameras of the vehicle to obtain a reference image and a target image, and to extract the overlap region of the reference image and the target image to obtain a reference overlap region image and a target overlap region image;

[0032] The segmentation module is used to segment the reference overlap area map to obtain a plurality of first overlap maps; segment the target overlap area map to obtain a plurality of second overlap maps; wherein each of the first overlap maps and each of the second overlap maps Figure 1 One to one correspondence;

[0033] The coincidence graph determination module is used to calculate similarity values ​​for each of the first coincidence graphs and the corresponding second coincidence graphs according to a preset algorithm, until the similarity value meets a preset threshold range, and use the first coincidence graph and the corresponding second coincidence graph whose similarity value meets the preset threshold range as the first determined coincidence graph and the second determined coincidence graph;

[0034] The processing module is used to process the target image according to the color features of the first determined coincidence image and the second determined coincidence image to complete color balancing processing.

[0035] Compared with the prior art, the above embodiment has the following beneficial effects: pre-processing the images taken by adjacent cameras to obtain a reference image and a target image, and extracting the overlapping area between them, thereby ensuring that subsequent processing is based on the content under the same perspective, which helps to improve the accuracy of color balancing, and by focusing on the overlapping area, the color differences between different cameras can be more accurately identified and adjusted; the reference overlapping area image and the target overlapping area image are respectively divided into multiple small area images, and the local color differences can be captured more finely through fine-grained segmentation, making the subsequent color correction more refined and accurate, and improving the sensitivity to subtle changes in complex scenes; by evaluating the similarity between the segmented images, the actual effective area that can be used for color balancing can be quickly screened out; by processing and adjusting the color features, the accuracy and consistency of the color balancing process are guaranteed.

[0036] In some embodiments of the second aspect of the present application, the coincidence graph determination module includes: a hash value calculation unit and a similarity value calculation unit;

[0037] The hash value calculation unit is used to calculate the image hash values ​​of the first coincidence image and the corresponding second coincidence image respectively;

[0038] The similarity value calculation unit is used to calculate the Hamming distance of the image hash value as the similarity value of the corresponding first coincidence image and the second coincidence image.

[0039] Compared with the prior art, the above embodiment has the following beneficial effects: by calculating the image hash value and the Hamming distance of the first coincidence image and the corresponding second coincidence image as the similarity value, it is possible to effectively determine whether the two coincidence images are from the same plane or scene, thereby ensuring the accuracy of subsequent color adjustment, avoiding the color distortion problem caused by non-coplanarity, and enhancing the robustness and adaptability of the algorithm.

[0040] In some embodiments of the second aspect of the present application, the processing module includes: an average value calculation unit, a difference calculation unit and a color difference equalization unit;

[0041] The average value calculation unit is used to respectively obtain the average values ​​of the first determined coincidence image and the second determined coincidence image on each of the three RGB channels;

[0042] The difference calculation unit is used to calculate the difference between the average values ​​of the first determined coincidence graph and the second determined coincidence graph on each channel according to the average values ​​on each channel;

[0043] The color difference equalization unit is used to process the target image according to the difference of the average values ​​on each channel to complete the color equalization processing.

[0044] Compared with the prior art, the above embodiment has the following beneficial effects: by obtaining and determining the average value of the overlapping images on the three RGB channels, calculating the difference and processing the target image accordingly, the image color can be accurately adjusted, so that the spliced ​​image is visually more natural and harmonious, and the color consistency of the image is improved.

[0045] In a third aspect, the present invention also provides a color balancing device for a vehicle's panoramic surround view image, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program, when loaded into the processor, implements the steps of the color balancing method for a vehicle's panoramic surround view image.

[0046] In a fourth aspect, an embodiment of the present application further provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the color balancing method of a vehicle panoramic surround view image are implemented. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 : A schematic flow chart of a color balancing method for a vehicle panoramic surround view image provided in some embodiments of the present invention.

[0048] Figure 2 : is a structural schematic diagram of a color balancing system for a vehicle panoramic surround view image provided in some embodiments of the present invention.

[0049] Figure 3 : is a structural diagram of a color balancing device for a vehicle panoramic surround view image provided in some embodiments of the present invention.

[0050] Figure 4 : A flowchart of extracting overlapping areas of a color balancing method for a vehicle panoramic surround view image provided in some embodiments of the present invention.

[0051] Figure 5 : An effect diagram of extracting overlapping areas of a color balancing method for a vehicle panoramic surround view image provided in some embodiments of the present invention.

[0052] Figure 6 : A schematic diagram of a segmentation overlap area diagram of a color balancing method for a vehicle panoramic surround view image provided in some embodiments of the present invention.

[0053] Figure 7 : An effect diagram of a segmentation overlap area diagram of a color balancing method for a vehicle panoramic surround view image provided in some embodiments of the present invention.

[0054] Figure 8 : A flowchart of image similarity judgment of a color balancing method for a vehicle panoramic surround view image provided in some embodiments of the present invention.

[0055] Fig. 9 : A flowchart of color equalization processing of a color equalization method for a vehicle panoramic surround view image provided in some embodiments of the present invention.

[0056] Fig.10 : is a diagram showing the actual effect of a color balancing method for a vehicle panoramic surround view image provided in some embodiments of the present invention.

[0057] Fig.11 : is a diagram showing the actual effect of a color balancing method for a vehicle panoramic surround view image provided in some embodiments of the present invention.

[0058] Fig.12 : is a diagram showing the actual effect of a color balancing method for a vehicle panoramic surround view image provided in some embodiments of the present invention.

[0059] Fig.13: An applicable scene diagram of a color balancing method for a vehicle panoramic surround view image provided in some embodiments of the present invention. DETAILED DESCRIPTION

[0060] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0061] Embodiment 1:

[0062] Please refer to Figure 1 , a color balancing method for a vehicle panoramic surround view image provided by an embodiment of the present invention, comprising steps S1 to S4:

[0063] Step S1: pre-processing the original images taken by the adjacent cameras of the vehicle to obtain a reference image and a target image, and extracting the overlapping area of ​​the reference image and the target image to obtain a reference overlapping area image and a target overlapping area image;

[0064] In some embodiments of the present application, step S1 can be implemented by the following preferred implementations:

[0065] S11: Acquire the original image captured by the camera adjacent to the vehicle;

[0066] S12: Processing each of the original images according to a preset calibration algorithm of the surround view system to obtain internal and external parameters and distortion coefficients of the camera;

[0067] S13: processing each of the original images according to a preset dedistortion algorithm of the surround view system and the distortion coefficient to obtain a dedistorted image;

[0068] S14: processing the dedistorted images according to the preset perspective transformation algorithm of the surround view system and the internal and external parameters to obtain top views corresponding to the dedistorted images, which are used as reference images and target images respectively;

[0069] S15: stitching the reference image and the target image according to a preset stitching algorithm of the surround view system to obtain a fused image;

[0070] S16: intercepting the overlapping area of ​​the fused image, taking the portion of the reference image corresponding to the overlapping area as a reference overlapping area map, and taking the portion of the target image corresponding to the overlapping area as a target overlapping area map.

[0071] For example, Figure 4A flowchart for extracting overlapping areas is shown. In a surround view system (AVM system), four images to be stitched are obtained by following basic steps such as calibration, image dedistortion, and perspective transformation, which are a front view, a rear view, a left view, and a right view. The images of two adjacent views will produce overlapping areas after stitching. There are four overlapping areas after the front, back, left, and right cameras are stitched together, corresponding to which a total of eight overlapping area images will be generated in the four images to be stitched. It is particularly noted that the calibration, dedistortion, perspective transformation, and stitching algorithms used in the surround view system (AVM system) in this application are all prior arts and will not be described in detail here. Of course, other similar extraction schemes can also be used for replacement and adjustment.

[0072] In addition, according to the different positions of the splicing seams, the size of the overlapped area can be defined according to actual needs. The present invention selects the right view and the rear view as examples, wherein the rear view is a reference view and the right view is a target view. Figure 5 The effect diagram of extracting the overlapping area is shown. According to the above steps, the overlapping area images of the two images to be stitched are extracted, that is, image A is the reference overlapping area image, and image B is the target overlapping area image.

[0073] In this embodiment, step S1 uses basic steps such as calibration, dedistortion and perspective transformation to process the original image, ensuring that the images captured by each camera can be accurately spliced ​​in the same coordinate system, laying a foundation for subsequent color balancing processing, and effectively solving the image differences caused by inconsistent camera installation heights, thereby improving the quality and reliability of image processing; by splicing the reference image and the target image to obtain a fused image, and intercepting the overlapping area therefrom, it is ensured that subsequent processing is performed on the actual overlapping part, which helps to improve the color balancing effect; in addition, this process can also flexibly define the size of the overlapping area to meet the needs of different application scenarios, thereby increasing the flexibility and practicality of the system.

[0074] Step S2: segmenting the reference overlap region map to obtain a plurality of first overlap regions; segmenting the target overlap region map to obtain a plurality of second overlap regions;

[0075] Wherein, each of the first overlapped graphs and each of the second overlapped graphs Figure 1 One to one correspondence.

[0076] In some embodiments of the present application, step S2 may be implemented by one of the following implementations:

[0077] For example, Figure 6 The schematic diagram of a segmentation overlap area map shown in FIG. can evenly divide image A and image B into 10 small module maps respectively, each small module is named in sequence from 1 to 10, and the 10 small module maps of image A and the 10 small modules of image B Figure 1 One to one correspondence.

[0078] In this embodiment, step S2 divides the reference overlap area map and the target overlap area map into multiple small area maps respectively. The fine-grained segmentation can capture the local color differences more finely, making the subsequent color correction more refined and accurate, and improving the sensitivity to subtle changes in complex scenes.

[0079] Step S3: Calculate similarity values ​​for each of the first coincidence graphs and the corresponding second coincidence graphs according to a preset algorithm, until the similarity values ​​meet a preset threshold range, and use the first coincidence graph and the corresponding second coincidence graph whose similarity values ​​meet the preset threshold range as the first determined coincidence graph and the second determined coincidence graph;

[0080] In some embodiments of the present application, step S3 can be implemented by the following preferred implementations:

[0081] S31: Calculate image hash values ​​of the first coincidence image and the corresponding second coincidence image respectively;

[0082] S32: Calculate the Hamming distance of the image hash value as the similarity value of the corresponding first coincidence graph and the second coincidence graph.

[0083] For example, Figure 7 As shown in the effect diagram of the segmentation overlap area diagram, starting from the first small module diagram, the small module diagrams with sequence number 1 of diagram A and diagram B are taken out as the data for the next step of processing; Figure 8 A flowchart of image similarity judgment is shown, in which two small module images with serial number 1 are processed in steps S31 and S32, and the similarity of the two small module images, namely the first overlap image and the second overlap image, is judged. If the judgment result is similar, the two small module images with serial number 1 are used as data input for subsequent processing. If they are judged to be dissimilar, the two small module images with serial number 2 are selected for similarity judgment. The method is the same as step S3, and so on, until two small module images whose similarity values ​​meet the preset conditions are found as image data for subsequent processing. If no similar image is found after traversing 10 small module area images, it means that the overlap area is not based on the same plane. If forced adjustment may cause color distortion of the picture, the adjustment is stopped.

[0084] In this embodiment, step S3 can effectively determine whether the two coincident images are from the same plane or scene by calculating the image hash value and the Hamming distance of the first coincident image and the corresponding second coincident image as the similarity value, thereby ensuring the accuracy of subsequent color adjustment, avoiding color distortion problems caused by non-coplanarity, and enhancing the robustness and adaptability of the algorithm.

[0085] In particular, for the panoramic view system, different cameras are installed at different heights, resulting in different captured images. When the captured objects are at a certain height, the images in the overlapping areas will be very different, such as Fig.13 An applicable scene diagram is shown, and the scene is a lawn with a height. Because the overall color difference between the two pictures is very large, if the entire overlapping area is directly applied to adjust the parameters, it will deviate far from the real color, resulting in a very poor picture balance effect. The algorithm of the present invention can not only consider the flat ground picture taken by the surround camera, but also optimize the scene with similar height difference, add small module image segmentation and similarity judgment technology, ensure that all parameter calculations are based on the same plane, and improve the consistency and accuracy of image color.

[0086] Step S4: processing the target image according to the color features of the first determined coincidence image and the second determined coincidence image to complete color balancing processing.

[0087] In some embodiments of the present application, Fig. 9 The flowchart of a color balancing process is shown, and step S4 can be implemented by the following preferred implementation:

[0088] S41: respectively obtaining average values ​​of the first determined coincidence map and the second determined coincidence map on each of the three RGB channels;

[0089] S42: Calculate the difference between the average values ​​of the first determined coincidence graph and the second determined coincidence graph on each channel according to the average values ​​on each channel;

[0090] S43: Process the target image according to the difference between the average values ​​on each channel to complete the color balancing process.

[0091] In this embodiment, step S4 obtains and determines the average value of the overlapped image on the RGB three channels, calculates the difference and processes the target image accordingly, so as to accurately adjust the image color, make the spliced ​​image more natural and harmonious visually, and improve the color consistency of the image.

[0092] In summary, after the processing of steps S1 to S4, Fig.10 , 11 As shown in the actual effect diagram of the color balancing method in Figure 12, the algorithm of the present invention is applicable to more scenes. Compared with the prior art which generally only performs color balancing on general driving roads, the algorithm of the present invention has a certain picture balancing effect for scenes with height differences such as near grass, steps, and scenes with ground reflections on rainy days, and has stronger environmental adaptability.

[0093] In summary, compared with the prior art, the embodiments of the present application have the following beneficial effects: images taken by adjacent cameras are preprocessed to obtain reference images and target images, and the overlapping areas between them are extracted, ensuring that subsequent processing is based on the content under the same perspective, which helps to improve the accuracy of color balancing. By focusing on the overlapping areas, the color differences between different cameras can be more accurately identified and adjusted; the reference overlapping area map and the target overlapping area map are respectively divided into multiple small area maps, and the local color differences can be captured more finely through fine-grained segmentation, making the subsequent color correction more refined and accurate, and improving the sensitivity to subtle changes in complex scenes; by evaluating the similarity between the segmented images, the actual effective area that can be used for color balancing can be quickly screened out; by processing and adjusting the color features, the accuracy and consistency of the color balancing process are guaranteed.

[0094] Embodiment 2:

[0095] Please refer to Figure 2 Based on the same inventive concept, an embodiment of the present invention discloses a color balancing system for a vehicle panoramic surround view image, comprising: an overlap region acquisition module M1, a segmentation module M2, an overlap map determination module M3 and a processing module M4.

[0096] The overlap region acquisition module M1 is used to pre-process the original images taken by the adjacent cameras of the vehicle to obtain a reference image and a target image, and to extract the overlap region of the reference image and the target image to obtain a reference overlap region image and a target overlap region image;

[0097] The overlap region acquisition module M1 includes: a collection unit, a calibration unit, a dedistortion unit, a perspective transformation unit, a splicing unit and a cutting unit;

[0098] Wherein, the acquisition unit is used to obtain the original image taken by the vehicle's adjacent camera;

[0099] The calibration unit is used to process each of the original images according to a preset calibration algorithm of the surround view system to obtain the internal and external parameters and distortion coefficients of the camera;

[0100] The dedistortion unit is used to process each of the original images according to a preset dedistortion algorithm of the surround view system and the distortion coefficient to obtain a dedistorted image;

[0101] The perspective transformation unit is used to process each of the dedistorted images according to a preset perspective transformation algorithm of the surround view system and the internal and external parameters to obtain each top view corresponding to the dedistorted image as a reference image and a target image respectively;

[0102] The stitching unit is used to stitch the reference image and the target image according to a preset stitching algorithm of the surround view system to obtain a fused image;

[0103] The interception unit is used to intercept the overlapping area of ​​the fused image, use the part of the reference image corresponding to the overlapping area as the reference overlapping area map, and use the part of the target image corresponding to the overlapping area as the target overlapping area map.

[0104] The overlapping area acquisition module M1 of this embodiment processes the original image by adopting basic steps such as calibration, dedistortion and perspective transformation, thereby ensuring that the images captured by each camera can be accurately spliced ​​in the same coordinate system, laying the foundation for subsequent color balancing processing, and effectively solving the image differences caused by inconsistent camera installation heights, thereby improving the quality and reliability of image processing; by splicing the reference image and the target image to obtain a fused image, and extracting the overlapping area therefrom, it ensures that subsequent processing is performed on the actual overlapping part, which helps to improve the color balancing effect; in addition, this process can also flexibly define the size of the overlapping area to meet the needs of different application scenarios, thereby increasing the flexibility and practicality of the system.

[0105] The segmentation module M2 is used to segment the reference overlap area map to obtain a plurality of first overlap maps; segment the target overlap area map to obtain a plurality of second overlap maps; wherein each of the first overlap maps and each of the second overlap maps Figure 1 One to one correspondence;

[0106] The segmentation module M2 of this embodiment divides the reference overlap area map and the target overlap area map into multiple small area maps respectively. Through fine-grained segmentation, it can capture local color differences more finely, making subsequent color correction more refined and accurate, and improving sensitivity to subtle changes in complex scenes.

[0107] The coincidence graph determination module M3 is used to calculate similarity values ​​for each of the first coincidence graphs and the corresponding second coincidence graphs according to a preset algorithm, until the similarity values ​​meet a preset threshold range, and use the first coincidence graph and the corresponding second coincidence graph whose similarity values ​​meet the preset threshold range as the first determined coincidence graph and the second determined coincidence graph;

[0108] The coincidence graph determination module M3 includes: a hash value calculation unit and a similarity value calculation unit;

[0109] The hash value calculation unit is used to calculate the image hash values ​​of the first coincidence image and the corresponding second coincidence image respectively;

[0110] The similarity value calculation unit is used to calculate the Hamming distance of the image hash value as the similarity value of the corresponding first coincidence image and the second coincidence image.

[0111] The overlap image determination module M3 of this embodiment can effectively determine whether the two overlap images are from the same plane or scene by calculating the image hash value and the Hamming distance of the first overlap image and the corresponding second overlap image as the similarity value, thereby ensuring the accuracy of subsequent color adjustment, avoiding color distortion problems caused by non-coplanarity, and enhancing the robustness and adaptability of the algorithm.

[0112] The processing module M4 is used to process the target image according to the color features of the first determined coincidence image and the second determined coincidence image to complete color balancing processing.

[0113] The processing module M4 includes: an average value calculation unit, a difference calculation unit and a color difference equalization unit;

[0114] The average value calculation unit is used to respectively obtain the average values ​​of the first determined coincidence image and the second determined coincidence image on each of the three RGB channels;

[0115] The difference calculation unit is used to calculate the difference between the average values ​​of the first determined coincidence graph and the second determined coincidence graph on each channel according to the average values ​​on each channel;

[0116] The color difference equalization unit is used to process the target image according to the difference of the average values ​​on each channel to complete the color equalization processing.

[0117] The processing module M4 of this embodiment can accurately adjust the image color by obtaining the average value of the overlapped image on the RGB three channels, calculating the difference and processing the target image accordingly, so that the spliced ​​image is more natural and harmonious visually, thereby improving the color consistency of the image.

[0118] In summary, compared with the prior art, the embodiments of the present application have the following beneficial effects: images taken by adjacent cameras are preprocessed to obtain reference images and target images, and the overlapping areas between them are extracted, ensuring that subsequent processing is based on the content under the same perspective, which helps to improve the accuracy of color balancing. By focusing on the overlapping areas, the color differences between different cameras can be more accurately identified and adjusted; the reference overlapping area map and the target overlapping area map are respectively divided into multiple small area maps, and the local color differences can be captured more finely through fine-grained segmentation, making the subsequent color correction more refined and accurate, and improving the sensitivity to subtle changes in complex scenes; by evaluating the similarity between the segmented images, the actual effective area that can be used for color balancing can be quickly screened out; by processing and adjusting the color features, the accuracy and consistency of the color balancing process are guaranteed.

[0119] Embodiment three:

[0120] Figure 3The structure diagram of a color balancing device for a vehicle panoramic surround image of the present application is shown. Figure 3 As shown, the color balancing device may include: a processor N1, a memory N2, a data interface N3 and a communication bus N4.

[0121] Among them: the processor N1, the memory N2, and the data interface N3 communicate with each other through the communication bus N4; the data interface N3 is used for data communication with other devices such as an input device or an output device; the processor N1 is used to execute the program N5, which can specifically execute the relevant steps in the above-mentioned embodiment of the color balancing method of the panoramic surround view image of the vehicle.

[0122] Specifically, the program N5 may include program codes, and the program codes include computer executable instructions.

[0123] The processor N1 may be a central processing unit CPU, or an application specific integrated circuit ASIC (Application Specific Integrated Circuit), or one or more integrated circuits configured to implement the embodiments of the present application. The one or more processors included in the color balancing device for the vehicle panoramic surround image may be processors of the same type, such as one or more CPUs, or may be processors of different types, such as one or more CPUs and one or more ASICs.

[0124] The memory N2 is used to store the program N5. The memory N2 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), such as at least one disk storage.

[0125] The algorithms or displays provided herein are not inherently related to any particular computer, virtual system or other device. In addition, the embodiments of the present application are not directed to any particular programming language.

[0126] Embodiment 4:

[0127] An embodiment of the present invention further provides a computer-readable storage medium storing at least one executable instruction. When the executable instruction is executed on a color balancing device / system for a vehicle panoramic surround view image, the color balancing device / system for a vehicle panoramic surround view image executes a color balancing method for a vehicle panoramic surround view image in any of the above method embodiments.

[0128] In the specification provided herein, a large number of specific details are described. However, it is understood that the embodiments of the present application can be practiced without these specific details. Similarly, in order to simplify the application and help understand one or more of the various inventive aspects, in the above description of the exemplary embodiments of the present application, the various features of the embodiments of the present application are sometimes grouped together into a single embodiment, figure, or description thereof. Wherein, the claims that follow the specific embodiment are thus clearly incorporated into the specific embodiment, wherein each claim itself is used as a separate embodiment of the application.

[0129] Those skilled in the art will appreciate that the modules in the devices in the embodiments may be adaptively changed and arranged in one or more devices different from the embodiments. The modules or units or components in the embodiments may be combined into one module or unit or component, and further may be divided into a plurality of submodules or subunits or subcomponents. Except that at least some of such features and / or processes or units are mutually exclusive.

Claims

1. A color balancing method for a vehicle panoramic surround image, characterized in that: include: Preprocessing the original images taken by the adjacent cameras of the vehicle to obtain a reference image and a target image, and extracting the overlapping area of ​​the reference image and the target image to obtain a reference overlapping area image and a target overlapping area image; Segmenting the reference coincidence region map to obtain a plurality of first coincidence maps; segmenting the target coincidence region map to obtain a plurality of second coincidence maps; wherein each of the first coincidence maps corresponds to each of the second coincidence maps one by one; According to a preset algorithm, respectively, a similarity value is calculated for each of the first coincidence graphs and the corresponding second coincidence graphs until the similarity value meets a preset threshold range, and the first coincidence graph and the corresponding second coincidence graph whose similarity value meets the preset threshold range are used as the first determined coincidence graph and the second determined coincidence graph; The target image is processed according to the color features of the first determined coincidence image and the second determined coincidence image to complete the color balancing process.

2. The color balancing method for a vehicle panoramic surround image according to claim 1, characterized in that: The calculating of similarity values ​​for each of the first coincidence graphs and the corresponding second coincidence graphs according to a preset algorithm includes: Calculating image hash values ​​of the first coincidence image and the corresponding second coincidence image respectively; The Hamming distance of the hash value of the image is calculated as a similarity value between the corresponding first coincidence image and the second coincidence image.

3. The color balancing method for a vehicle panoramic surround image as claimed in claim 2, characterized in that: The processing of the target image according to the color features of the first determined coincidence image and the second determined coincidence image to complete the color balancing process includes: Respectively obtain the average values ​​of the first determined coincidence map and the second determined coincidence map on each channel of the RGB three channels; Calculate the difference between the average values ​​of the first determined coincidence graph and the second determined coincidence graph on each channel according to the average values ​​on each channel; The target image is processed according to the difference between the average values ​​on each channel to complete the color balancing process.

4. A color balancing method for a vehicle panoramic view image according to any one of claims 1 to 3, characterized in that: The preprocessing of the original images captured by the adjacent cameras of the vehicle to obtain the reference image and the target image includes: Obtain the original image taken by the vehicle's adjacent cameras; According to a preset calibration algorithm of the surround view system, each of the original images is processed to obtain the internal and external parameters and distortion coefficients of the camera; Processing each of the original images according to a preset dedistortion algorithm of the surround view system and the distortion coefficient to obtain a dedistorted image; According to the preset perspective transformation algorithm of the surround view system and the internal and external parameters, each of the dedistorted images is processed to obtain each top view corresponding to the dedistorted image, which is used as a reference image and a target image respectively.

5. The color balancing method for a vehicle panoramic view image as claimed in claim 4, characterized in that: The step of extracting the overlapped area of ​​the reference image and the target image to obtain a reference overlapped area map and a target overlapped area map includes: splicing the reference image and the target image according to a preset stitching algorithm of the surround view system to obtain a fused image; The overlapping area of ​​the fused image is intercepted, and the portion of the reference image corresponding to the overlapping area is used as a reference overlapping area map, and the portion of the target image corresponding to the overlapping area is used as a target overlapping area map.

6. A color balancing system for a vehicle panoramic surround image, characterized in that: include: Overlapping area acquisition module, segmentation module, overlap map determination module and processing module; The overlap region acquisition module is used to pre-process the original images taken by the adjacent cameras of the vehicle to obtain a reference image and a target image, and to extract the overlap region of the reference image and the target image to obtain a reference overlap region image and a target overlap region image; The segmentation module is used to segment the reference coincidence area map to obtain a plurality of first coincidence maps; segment the target coincidence area map to obtain a plurality of second coincidence maps; wherein each of the first coincidence maps corresponds to each of the second coincidence maps one by one; The coincidence graph determination module is used to calculate similarity values ​​for each of the first coincidence graphs and the corresponding second coincidence graphs according to a preset algorithm, until the similarity value meets a preset threshold range, and use the first coincidence graph and the corresponding second coincidence graph whose similarity value meets the preset threshold range as the first determined coincidence graph and the second determined coincidence graph; The processing module is used to process the target image according to the color features of the first determined coincidence image and the second determined coincidence image to complete color balancing processing.

7. A color balancing system for vehicle panoramic surround view images as claimed in claim 6, characterized in that: The coincidence graph determination module includes: a hash value calculation unit and a similarity value calculation unit; The hash value calculation unit is used to calculate the image hash values ​​of the first coincidence image and the corresponding second coincidence image respectively; The similarity value calculation unit is used to calculate the Hamming distance of the image hash value as the similarity value of the corresponding first coincidence image and the second coincidence image.

8. The color balancing system for vehicle panoramic surround view images as claimed in claim 7, characterized in that: The processing module includes: an average value calculation unit, a difference calculation unit and a color difference equalization unit; The average value calculation unit is used to respectively obtain the average values ​​of the first determined coincidence image and the second determined coincidence image on each of the three RGB channels; The difference calculation unit is used to calculate the difference between the average values ​​of the first determined coincidence graph and the second determined coincidence graph on each channel according to the average values ​​on each channel; The color difference equalization unit is used to process the target image according to the difference of the average values ​​on each channel to complete the color equalization processing.

9. An electronic device for color balancing of a vehicle's panoramic surround image, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the computer program is loaded into a processor, the steps of the color balancing method for a vehicle panoramic surround view image according to any one of claims 1 to 5 are implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the color balancing method for a vehicle panoramic surround view image according to any one of claims 1 to 5 are implemented.

Images