Image processing method, device, equipment and storage medium based on artificial intelligence

Abstract

The present invention relates to the field of image processing technology, and discloses an image processing method, device, equipment and storage medium based on artificial intelligence. By acquiring the traffic monitoring image of each traffic camera in the target area, performing a first color conversion using regional environmental information, a first optimized monitoring image is obtained, and then based on the vehicle traffic image corresponding to each vehicle traffic action in the initial matching period of each monitoring cycle, an image allocation reference data set is constructed, so as to realize the task allocation and second color conversion of the first optimized monitoring image with respect to the image processing terminal, and obtain the final second optimized monitoring image. Thus, the color directional distortion caused by environmental factors and the color non-directional distortion caused by some sensor failures are considered, the color restoration efficiency and accuracy of the traffic monitoring image are improved, and the number of processing of the vehicle traffic image is refined when performing the color restoration task based on the traffic monitoring image, so as to realize the reasonable allocation and efficient processing of the color restoration task.

Description

Technical Field

[0001] The present invention relates to the field of image processing technology, and in particular to an image processing method, device, equipment and storage medium based on artificial intelligence. Background Art

[0002] Traffic cameras are video cameras installed on roads to monitor traffic conditions and record traffic violations. They play an important role in modern urban traffic management and are an important tool for ensuring traffic order and safety. The video images collected by traffic cameras are usually transmitted to image processing terminals, and traffic conditions and traffic violations are identified through video image analysis technology. Therefore, the image quality of video images (parameters such as clarity and color reproduction) is usually highly required.

[0003] However, in actual use, the video image acquisition and processing of traffic cameras have the following problems: First, the color distortion of video images leads to a decrease in image recognition accuracy, where the color distortion of video images is caused by multiple aspects, including directional distortion of image color caused by environmental factors and non-directional distortion of image color caused by partial sensor failure of the camera. The existing video image color restoration method for traffic cameras does not have a good solution for the color distortion caused by a combination of multiple factors; second, due to the time period differences and stages of road traffic, the image processing task volume of video images collected by traffic cameras at different locations at different times is different and variable, especially when it comes to the need to restore the color distortion of video images collected by traffic cameras, which has high requirements on the image processing capabilities of the image processing terminal, the fixed allocation strategy between traditional traffic cameras and image processing terminals will not be directly applicable.

[0004] Therefore, how to restore the color of traffic monitoring images collected by traffic cameras that are distorted by a combination of multiple factors and improve the rationality of task allocation and processing efficiency of traffic cameras when performing color restoration tasks is a technical problem that needs to be solved urgently. Summary of the invention

[0005] The main purpose of the present invention is to provide an image processing method, device, equipment and storage medium based on artificial intelligence to solve the above-mentioned existing technical problems.

[0006] To achieve the above object, the present invention provides an image processing method based on artificial intelligence, comprising the following steps:

[0007] Acquire traffic monitoring information of the target area; wherein the traffic monitoring information includes identification information of a plurality of traffic cameras and a traffic monitoring image captured by each traffic camera;

[0008] Based on the identification information of each traffic camera, query the regional environment information of each traffic camera;

[0009] Using the regional environment information, performing a first color conversion on the traffic monitoring image captured by each traffic camera to obtain a first optimized monitoring image;

[0010] In the initial matching period of each monitoring cycle, identifying the vehicle passage action in the first optimized monitoring image, extracting a plurality of vehicle passage images corresponding to each vehicle passage action in the initial matching period, and constructing an image allocation reference data set;

[0011] According to the image allocation reference data set, a matching relationship between the traffic camera and the image processing terminal is established;

[0012] The first optimized monitoring image in the image processing period of each monitoring cycle is sent to an image processing terminal having a matching relationship, and a second color conversion is performed to obtain a second optimized monitoring image.

[0013] Optionally, based on the identification information of each traffic camera, the step of querying the regional environment information of each traffic camera specifically includes:

[0014] Based on the identification information of each traffic camera, query the location information of each traffic camera in the traffic camera deployment data list;

[0015] The location information is used to access the regional environment database to extract the current regional environment information of each traffic camera; wherein the regional environment information includes a plurality of regional environment data that affect the color recognition of the traffic monitoring image.

[0016] Optionally, using the regional environment information, performing a first color conversion on the traffic monitoring image captured by each traffic camera to obtain a first optimized monitoring image step specifically includes:

[0017] The regional environment information is used as a regional environment feature set, and input into a pre-trained image color environment optimization model to obtain output image color environment optimization parameters for each traffic camera;

[0018] Wherein, the image color environment optimization parameter is configured as a numerical adjustment ratio of the RGB color value of each color channel;

[0019] Based on the image color environment optimization parameters, a first color conversion is performed on the RGB color value of each pixel in the traffic monitoring image captured by each camera to obtain a first optimized monitoring image;

[0020] Among them, the first color conversion is configured to multiply the RGB color value of each pixel in the traffic monitoring image by the numerical adjustment ratio of the corresponding color channel, and use the multiplication result as the RGB color value of the corresponding pixel in the first optimized monitoring image.

[0021] Optionally, before the step of using the regional environment information as a regional environment feature set and inputting a pre-trained image color environment optimization model to obtain output image color environment optimization parameters for each traffic camera, the method further includes:

[0022] Acquire test color images of standard color sample images collected by traffic cameras under different regional environmental feature combinations; wherein the regional environmental feature combination includes one or more of the following: solar altitude angle, atmospheric component content, cloud thickness parameter, and street light illumination color;

[0023] A training sample set is constructed based on the standard color value of each standard color block in the standard color sample image, the combination of environmental features of different regions, and the photographed color value of each standard color block in the test color image;

[0024] The training sample set is input into the constructed initial neural network model for training, and an image color environment optimization model is obtained after the training times reaches a preset number or the training converges.

[0025] Optionally, in the initial matching period of each monitoring cycle, identifying the vehicle passage action in the first optimized monitoring image, extracting a plurality of vehicle passage images corresponding to each vehicle passage action in the initial matching period, and constructing the image allocation reference data set step specifically includes:

[0026] In the initial matching period of each monitoring cycle, a vehicle detection algorithm is used to identify the vehicle passing action in the first optimized monitoring image, and a passing vehicle anchor frame is generated in the first optimized monitoring image;

[0027] Based on the passing vehicle anchor frame, a plurality of vehicle passing images corresponding to each vehicle passing action within the initial matching period are extracted, and the plurality of vehicle passing images are used as image allocation reference data sets for each vehicle passing action.

[0028] Optionally, the step of establishing a matching relationship between a traffic camera and an image processing terminal according to the image allocation reference data set specifically includes:

[0029] Identify the target vehicle traffic image in the image allocation reference data set for each vehicle traffic action of each traffic camera in the initial matching period of each monitoring cycle, and generate the hardware optimization standard coefficient of each traffic camera;

[0030] The target vehicle traffic image is configured as an image whose vehicle RGB color value in the image allocation reference data set is abnormal compared with the vehicle RGB color value of the non-target vehicle traffic image in the image allocation reference data set, and the proportion of the target vehicle traffic image in the image allocation reference data set is smaller than the proportion of the non-target vehicle traffic image in the image allocation reference data set;

[0031] The hardware optimization standard coefficient includes a plurality of optimized pixel points with RGB color value adjustment values, the optimized pixel points are configured as a set of pixel points with abnormal vehicle RGB color values ​​in all target vehicle images, and the RGB color value adjustment value is configured as an average value of the RGB color difference between the target vehicle passing image and the non-target vehicle passing image at the same position of the vehicle with respect to each color channel in each vehicle passing action in the initial matching period;

[0032] Calculate the number of target vehicle traffic images for all vehicle traffic actions in the initial matching period of each monitoring cycle as the image processing requirement of each traffic camera;

[0033] Based on the image processing requirements of each traffic camera and the image processing capabilities of the image processing terminal, the traffic camera and the image processing terminal are matched to establish a matching relationship between the traffic camera and the image processing terminal;

[0034] The image processing capability of the image processing terminal is configured as the number of standard test pictures processed by the image processing terminal within a unit time.

[0035] Optionally, the step of sending the first optimized monitoring image in the image processing period of each monitoring cycle to an image processing terminal having a matching relationship, performing a second color conversion, and obtaining a second optimized monitoring image specifically includes:

[0036] Sending the first optimized monitoring image and the hardware optimization standard coefficient of each traffic camera in the image processing period of each monitoring cycle to an image processing terminal having a matching relationship;

[0037] Each image processing terminal is driven to perform a second color conversion on the first optimized monitoring image according to the received hardware optimization standard coefficient to obtain a second optimized monitoring image.

[0038] In addition, in order to achieve the above-mentioned purpose, the present invention also provides an image processing device based on artificial intelligence, comprising:

[0039] An acquisition module, used to acquire traffic monitoring information of a target area; wherein the traffic monitoring information includes identification information of a plurality of traffic cameras and a traffic monitoring image captured by each traffic camera;

[0040] A query module, used to query the regional environment information of each traffic camera based on the identification information of each traffic camera;

[0041] A first execution module, configured to perform a first color conversion on the traffic monitoring image captured by each traffic camera using the regional environment information to obtain a first optimized monitoring image;

[0042] An identification module is used to identify the vehicle passage action in the first optimized monitoring image during the initial matching period of each monitoring cycle, extract a number of vehicle passage images corresponding to each vehicle passage action during the initial matching period, and construct an image allocation reference data set;

[0043] Establishing a module for allocating reference data sets according to images and establishing a matching relationship between traffic cameras and image processing terminals;

[0044] The second execution module is used to send the first optimized monitoring image in the image processing period of each monitoring cycle to an image processing terminal with a matching relationship, perform a second color conversion, and obtain a second optimized monitoring image.

[0045] In addition, in order to achieve the above-mentioned purpose, the present invention also provides an artificial intelligence-based image processing device, which includes: a memory, a processor, and an artificial intelligence-based image processing program stored in the memory and executable on the processor, and the artificial intelligence-based image processing program, when executed by the processor, implements the steps of the artificial intelligence-based image processing method as described above.

[0046] In addition, in order to achieve the above-mentioned purpose, the present invention also provides an artificial intelligence-based storage medium, on which an artificial intelligence-based image processing program is stored, and when the artificial intelligence-based image processing program is executed by a processor, the steps of the above-mentioned artificial intelligence-based image processing method are implemented.

[0047] The beneficial effects of the present invention are as follows: an image processing method, device, equipment and storage medium based on artificial intelligence are proposed, by acquiring the traffic monitoring image and regional environmental information of each traffic camera in the target area, using the regional environmental information to perform a first color conversion on the traffic monitoring image, obtaining a first optimized monitoring image, and then based on the vehicle traffic image corresponding to each vehicle traffic action in the initial matching period of each monitoring cycle, constructing an image allocation reference data set and a hardware optimization standard coefficient, thereby realizing the task allocation of the first optimized monitoring image to the image processing terminal and the second color conversion of the hardware optimization, and obtaining the final second optimized monitoring image. Thus, the color directional distortion caused by environmental factors and the color non-directional distortion caused by some sensor failures are considered, the color restoration efficiency and accuracy of the traffic monitoring image are improved, and the number of processing of the vehicle traffic image is refined when performing the color restoration task based on the traffic monitoring image, so as to realize the reasonable allocation and efficient processing of the color restoration task. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 A schematic diagram of the device structure of the hardware operating environment involved in the embodiment of the present invention;

[0049] Figure 2 A schematic diagram of a flow chart of an embodiment of an image processing method based on artificial intelligence of the present invention;

[0050] Figure 3 4 is a structural block diagram of an image processing device based on artificial intelligence in an embodiment of the present invention.

[0051] The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0052] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0053] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0054] like Figure 1 As shown, Figure 1 It is a schematic diagram of the device structure of the hardware operating environment involved in the embodiment of the present invention.

[0055] like Figure 1As shown, the device may include: a processor 1001, such as a CPU, a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. Among them, the communication bus 1002 is used to realize the connection and communication between these components. The user interface 1003 may include a display screen (Display), an input unit such as a keyboard (Keyboard), and the optional user interface 1003 may also include a standard wired interface and a wireless interface. The network interface 1004 may optionally include a standard wired interface and a wireless interface (such as a WI-FI interface). The memory 1005 may be a high-speed RAM memory, or it may be a stable memory (non-volatile memory), such as a disk memory. The memory 1005 may also be a storage device independent of the aforementioned processor 1001.

[0056] Those skilled in the art will understand that Figure 1 The structure of the device shown in the figure does not constitute a limitation of the device, and may include more or less components than shown in the figure, or combine certain components, or arrange the components differently.

[0057] like Figure 1 As shown, the memory 1005 as a storage medium based on artificial intelligence may include an operating system, a network communication module, a user interface module, and an image processing program based on artificial intelligence.

[0058] exist Figure 1 In the terminal shown, the network interface 1004 is mainly used to connect to the backend server and communicate data with the backend server; the user interface 1003 is mainly used to connect to the client (user end) and communicate data with the client; and the processor 1001 can be used to call the artificial intelligence-based image processing program stored in the memory 1005 and perform the following operations:

[0059] Acquire traffic monitoring information of the target area; wherein the traffic monitoring information includes identification information of a plurality of traffic cameras and a traffic monitoring image captured by each traffic camera;

[0060] Based on the identification information of each traffic camera, query the regional environment information of each traffic camera;

[0061] Using the regional environment information, performing a first color conversion on the traffic monitoring image captured by each traffic camera to obtain a first optimized monitoring image;

[0062] In the initial matching period of each monitoring cycle, identifying the vehicle passage action in the first optimized monitoring image, extracting a plurality of vehicle passage images corresponding to each vehicle passage action in the initial matching period, and constructing an image allocation reference data set;

[0063] According to the image allocation reference data set, a matching relationship between the traffic camera and the image processing terminal is established;

[0064] The first optimized monitoring image in the image processing period of each monitoring cycle is sent to an image processing terminal having a matching relationship, and a second color conversion is performed to obtain a second optimized monitoring image.

[0065] The specific embodiments of the present invention applied to the device are basically the same as the embodiments of the following application of the artificial intelligence-based image processing method, and will not be described in detail here.

[0066] The embodiment of the present invention provides an image processing method based on artificial intelligence, referring to Figure 2 , Figure 2 The figure is a flowchart of an embodiment of an image processing method based on artificial intelligence of the present invention.

[0067] In this embodiment, the image processing method based on artificial intelligence includes the following steps:

[0068] S100: Acquire traffic monitoring information of a target area; wherein the traffic monitoring information includes identification information of a plurality of traffic cameras and a traffic monitoring image captured by each traffic camera;

[0069] S200: Based on the identification information of each traffic camera, query the regional environment information of each traffic camera;

[0070] S300: Using the regional environment information, performing a first color conversion on the traffic monitoring image captured by each traffic camera to obtain a first optimized monitoring image;

[0071] S400: In an initial matching period of each monitoring cycle, identifying a vehicle passage action in a first optimized monitoring image, extracting a plurality of vehicle passage images corresponding to each vehicle passage action in the initial matching period, and constructing an image allocation reference data set;

[0072] S500: establishing a matching relationship between a traffic camera and an image processing terminal according to an image allocation reference data set;

[0073] S600: Send the first optimized monitoring image in the image processing period of each monitoring cycle to an image processing terminal having a matching relationship, perform a second color conversion, and obtain a second optimized monitoring image.

[0074] It should be noted that in actual use, the video image acquisition and processing of traffic cameras have the following problems: First, the color distortion of video images leads to a decrease in image recognition accuracy, where the color distortion of video images is caused by multiple aspects, including directional distortion of image color caused by environmental factors and non-directional distortion of image color caused by partial sensor failure of the camera. The existing video image color restoration method for traffic cameras does not have a good solution for the color distortion caused by a combination of multiple factors; second, due to the time period differences and stages of road traffic, the image processing task volume of video images collected by traffic cameras at different locations at different times is different and variable, especially when it comes to the need to restore the color distortion of video images collected by traffic cameras, which has high requirements on the image processing capabilities of the image processing terminal, the fixed allocation strategy between traditional traffic cameras and image processing terminals will not be directly applicable.

[0075] In order to solve the above problems, this embodiment obtains the traffic monitoring image and regional environmental information of each traffic camera in the target area, performs a first color conversion on the traffic monitoring image using the regional environmental information, obtains a first optimized monitoring image, and then constructs an image allocation reference data set and a hardware optimization standard coefficient based on the vehicle traffic image corresponding to each vehicle traffic action in the initial matching period of each monitoring cycle, thereby realizing the task allocation of the first optimized monitoring image to the image processing terminal and the second color conversion of the hardware optimization, and obtaining the final second optimized monitoring image. Thus, the color directional distortion caused by environmental factors and the color non-directional distortion caused by some sensor failures are considered, the color restoration efficiency and accuracy of the traffic monitoring image are improved, and the number of processing of the vehicle traffic image is refined when performing the color restoration task based on the traffic monitoring image, so as to realize the reasonable allocation and efficient processing of the color restoration task.

[0076] In a preferred embodiment, based on the identification information of each traffic camera, the step of querying the regional environment information of each traffic camera specifically includes:

[0077] S210: Based on the identification information of each traffic camera, query the location information of each traffic camera in the traffic camera deployment data list;

[0078] S220: Using the location information, access the regional environment database to extract the current regional environment information of each traffic camera; wherein the regional environment information includes a plurality of regional environment data that affect the color recognition of the traffic monitoring image.

[0079] In this embodiment, the regional environment information of each traffic camera is queried, and the location is queried in the traffic camera deployment data list through the identification information of each traffic camera, and then the regional environment information is extracted from the regional environment database based on the location. It should be noted that the construction of the regional environment database can be achieved by calling an external storage server for each item of regional environment data.

[0080] In a preferred embodiment, the step of performing a first color conversion on the traffic monitoring image captured by each traffic camera using the regional environment information to obtain a first optimized monitoring image specifically includes:

[0081] S340: using the regional environment information as a regional environment feature set, inputting a pre-trained image color environment optimization model, and obtaining output image color environment optimization parameters of each traffic camera; wherein the image color environment optimization parameters are configured as numerical adjustment ratios of RGB color values ​​of each color channel;

[0082] S350: Based on the image color environment optimization parameter, perform a first color conversion on the RGB color value of each pixel in the traffic monitoring image captured by each camera to obtain a first optimized monitoring image;

[0083] Among them, the first color conversion is configured to multiply the RGB color value of each pixel in the traffic monitoring image by the numerical adjustment ratio of the corresponding color channel, and use the multiplication result as the RGB color value of the corresponding pixel in the first optimized monitoring image.

[0084] On this basis, the regional environment information is used as a regional environment feature set, input into a pre-trained image color environment optimization model, and before the step of obtaining the output image color environment optimization parameters of each traffic camera, the method further includes:

[0085] S310: Acquire a test color image of a standard color sample image collected by a traffic camera under different regional environmental feature combinations; wherein the regional environmental feature combination includes one or more of a solar altitude angle, an atmospheric component content, a cloud thickness parameter, and a street light illumination color;

[0086] S320: constructing a training sample set based on the standard color value of each standard color block in the standard color sample image, the combination of environmental features of different regions, and the photographed color value of each standard color block in the test color image;

[0087] S330: Inputting the training sample set into the constructed initial neural network model for training, and obtaining an image color environment optimization model after the training times reach a preset number or the training converges.

[0088] Taking into account the directional distortion of image color caused by environmental factors, that is, different solar altitude angles, different atmospheric components, clouds and atmospheric pollution will cause differences in illumination color. For example, under bluish cold light, the image may appear to have too much blue component, while under yellowish warm light, the image may have an increased yellow component (yellow is a mixture of red and green, which indirectly affects the RGB ratio). In this embodiment, in order to solve the directional distortion of image color caused by environmental factors, an image color environment optimization model based on a neural network architecture is constructed, and the image color environment optimization model is used to obtain the image color environment optimization parameters of each current traffic camera according to the regional environmental feature set collected in real time. Finally, the image color environment optimization parameters are used to adjust the numerical adjustment ratio of each color channel in the current RGB color value of each traffic camera to obtain the first optimized monitoring image.

[0089] In a preferred embodiment, in the initial matching period of each monitoring cycle, identifying the vehicle passage action in the first optimized monitoring image, extracting a plurality of vehicle passage images corresponding to each vehicle passage action in the initial matching period, and constructing an image allocation reference data set step specifically includes:

[0090] S410: In the initial matching period of each monitoring cycle, a vehicle detection algorithm is used to identify the vehicle passing action in the first optimized monitoring image, and a passing vehicle anchor frame is generated in the first optimized monitoring image;

[0091] S420: Based on the passing vehicle anchor frame, extract a plurality of vehicle passing images corresponding to each vehicle passing action within the initial matching period, and use the plurality of vehicle passing images as image allocation reference data sets for each vehicle passing action.

[0092] In this embodiment, by adopting the existing vehicle detection algorithm, the vehicle passing action is identified in the first optimized monitoring image, the passing vehicle anchor frame is generated, and a number of vehicle passing images are extracted as the image allocation reference data set for each vehicle passing action. It should be noted that the image allocation reference data set includes the vehicle passing actions collected by each traffic camera during the initial matching period. By subsequently analyzing the number of vehicle passing actions and the number of images that need to be color restored, it is possible to achieve a reasonable allocation of the first optimized monitoring image among the image processing terminals.

[0093] In a preferred embodiment, the step of establishing a matching relationship between a traffic camera and an image processing terminal according to an image allocation reference data set specifically includes:

[0094] S510: Identify the target vehicle traffic image in the image allocation reference data set for each vehicle traffic action of each traffic camera in the initial matching period of each monitoring cycle, and generate the hardware optimization standard coefficient of each traffic camera;

[0095] The target vehicle traffic image is configured as an image whose vehicle RGB color value in the image allocation reference data set is abnormal compared with the vehicle RGB color value of the non-target vehicle traffic image in the image allocation reference data set, and the proportion of the target vehicle traffic image in the image allocation reference data set is smaller than the proportion of the non-target vehicle traffic image in the image allocation reference data set;

[0096] The hardware optimization standard coefficient includes a plurality of optimized pixel points with RGB color value adjustment values, the optimized pixel points are configured as a set of pixel points with abnormal vehicle RGB color values ​​in all target vehicle images, and the RGB color value adjustment value is configured as an average value of the RGB color difference between the target vehicle passing image and the non-target vehicle passing image at the same position of the vehicle with respect to each color channel in each vehicle passing action in the initial matching period;

[0097] S520: Calculate the number of target vehicle traffic images for all vehicle traffic actions in the initial matching period of each monitoring cycle as the image processing requirement of each traffic camera;

[0098] S530: Based on the image processing requirements of each traffic camera and the image processing capability of the image processing terminal, matching the traffic camera with the image processing terminal is performed to establish a matching relationship between the traffic camera and the image processing terminal;

[0099] The image processing capability of the image processing terminal is configured as the number of standard test pictures processed by the image processing terminal within a unit time.

[0100] In this embodiment, when using the image allocation reference data set to allocate the first optimized monitoring image between the image processing terminals, the number of target vehicle traffic images (i.e., images with abnormal vehicle RGB color values) of all vehicle traffic actions of each traffic camera in the initial matching period of each monitoring cycle is considered, and the traffic camera and the image processing terminal are matched according to the number and the image processing capability of the image processing terminal (for example, the number is sorted according to the image processing capability and then the corresponding allocation is performed according to the sequence number). At the same time, in this step, the hardware optimization standard coefficient of each traffic camera can also be generated according to the average value of the RGB color difference between the target vehicle traffic image and the non-target vehicle traffic image at the same position of the vehicle in each vehicle traffic action in the initial matching period, that is, it can be known which pixels in the current traffic camera have abnormal color values ​​due to sensor failure, and the RGB color value adjustment value for each pixel for color restoration reference.

[0101] In a preferred embodiment, the first optimized monitoring image in the image processing period of each monitoring cycle is sent to an image processing terminal having a matching relationship, and a second color conversion is performed to obtain a second optimized monitoring image, which specifically includes:

[0102] S610: Sending the first optimized monitoring image and the hardware optimization standard coefficient of each traffic camera in the image processing period of each monitoring cycle to an image processing terminal having a matching relationship;

[0103] S620: driving each image processing terminal to perform a second color conversion on the first optimized monitoring image according to the received hardware optimization standard coefficient to obtain a second optimized monitoring image.

[0104] In this embodiment, after obtaining the matching relationship between the traffic camera and the image processing terminal, the first optimized monitoring image and the hardware optimization standard coefficient of the image processing period are sent to the image processing terminal with the matching relationship, so as to realize the second color conversion of the first optimized monitoring image and obtain the second optimized monitoring image. It should be noted that the initial matching period of each monitoring cycle is the period of the preset proportion of time before each monitoring cycle, and the image processing period is the period after the initial matching period of each monitoring cycle (for example, each monitoring cycle is three hours, and the initial matching period is 1 minute). In some scenarios, a dynamically changing monitoring period can be set to adapt to the monitoring area with large changes in traffic flow, which can improve the scene adaptability of image color restoration processing.

[0105] This embodiment provides an image processing method based on artificial intelligence, which obtains the traffic monitoring image and regional environmental information of each traffic camera in the target area, performs a first color conversion on the traffic monitoring image using the regional environmental information, obtains a first optimized monitoring image, and then constructs an image allocation reference data set and a hardware optimization standard coefficient based on the vehicle traffic image corresponding to each vehicle traffic action in the initial matching period of each monitoring cycle, thereby realizing the task allocation of the first optimized monitoring image to the image processing terminal and the second color conversion of the hardware optimization, and obtaining the final second optimized monitoring image. Thus, the color directional distortion caused by environmental factors and the color non-directional distortion caused by some sensor failures are considered, the color restoration efficiency and accuracy of the traffic monitoring image are improved, and the number of processing of the vehicle traffic image is refined when performing the color restoration task based on the traffic monitoring image, so as to realize the reasonable allocation and efficient processing of the color restoration task.

[0106] Reference Figure 3 , Figure 3 It is a structural block diagram of an embodiment of an image processing device based on artificial intelligence of the present invention.

[0107] like Figure 3 As shown, the image processing device based on artificial intelligence proposed in the embodiment of the present invention includes:

[0108] The acquisition module 10 is used to acquire traffic monitoring information of the target area; wherein the traffic monitoring information includes identification information of a plurality of traffic cameras and a traffic monitoring image captured by each traffic camera;

[0109] A query module 20, for querying regional environment information of each traffic camera based on identification information of each traffic camera;

[0110] A first execution module 30 is used to perform a first color conversion on the traffic monitoring image captured by each traffic camera using the regional environment information to obtain a first optimized monitoring image;

[0111] The recognition module 40 is used to recognize the vehicle passage action in the first optimized monitoring image during the initial matching period of each monitoring cycle, extract a plurality of vehicle passage images corresponding to each vehicle passage action during the initial matching period, and construct an image allocation reference data set;

[0112] Establishing module 50, for establishing a matching relationship between a traffic camera and an image processing terminal according to an image allocation reference data set;

[0113] The second execution module 60 is used to send the first optimized monitoring image in the image processing period of each monitoring cycle to the image processing terminal with a matching relationship, perform a second color conversion, and obtain a second optimized monitoring image.

[0114] Other embodiments or specific implementations of the artificial intelligence-based image processing device of the present invention can refer to the above-mentioned method embodiments and will not be described in detail here.

[0115] In addition, the present invention also proposes an artificial intelligence-based image processing device, which includes: a memory, a processor, and an artificial intelligence-based image processing program stored in the memory and executable on the processor, and the artificial intelligence-based image processing program, when executed by the processor, implements the steps of the artificial intelligence-based image processing method as described above.

[0116] The specific implementation of the artificial intelligence-based image processing device of the present invention is basically the same as the above-mentioned embodiments of the artificial intelligence-based image processing method, and will not be repeated here.

[0117] In addition, the present invention also proposes an artificial intelligence-based storage medium, wherein the artificial intelligence-based storage medium comprises a computer-readable storage medium on which an artificial intelligence-based image processing program is stored. The artificial intelligence-based storage medium may be Figure 1 The memory 1005 in the terminal may also be at least one of a ROM (Read-Only Memory) / RAM (Random Access Memory), a magnetic disk, and an optical disk. The artificial intelligence-based storage medium includes a number of instructions for enabling an artificial intelligence-based image processing device with a processor to execute the artificial intelligence-based image processing method described in each embodiment of the present invention.

[0118] The specific implementation of the artificial intelligence-based storage medium of the present invention is basically the same as the above-mentioned embodiments of the artificial intelligence-based image processing method, and will not be repeated here.

[0119] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in an artificial intelligence-based storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes a number of instructions for a terminal device (which can be a mobile phone, computer, server, or network device, etc.) to execute the methods described in each embodiment of the present invention.

[0120] The above are only preferred embodiments of the present invention, and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. An image processing method based on artificial intelligence, It is characterized in that The following steps are involved: Acquire traffic monitoring information of the target area; wherein the traffic monitoring information includes identification information of a plurality of traffic cameras and a traffic monitoring image captured by each traffic camera; Based on the identification information of each traffic camera, query the regional environment information of each traffic camera; Utilizing the regional environment information, a first color conversion is performed on the traffic monitoring image taken by each traffic camera to obtain a first optimized monitoring image; specifically comprising: using the regional environment information as a regional environment feature set, inputting a pre-trained image color environment optimization model, and obtaining output image color environment optimization parameters for each traffic camera; wherein the image color environment optimization parameters are configured as a numerical adjustment ratio of the RGB color value of each color channel; based on the image color environment optimization parameters, a first color conversion is performed on the RGB color value of each pixel in the traffic monitoring image taken by each camera to obtain a first optimized monitoring image; wherein the first color conversion is configured to multiply the RGB color value of each pixel in the traffic monitoring image by the numerical adjustment ratio of the corresponding color channel, and using the multiplication result as the RGB color value of the corresponding pixel in the first optimized monitoring image; In the initial matching period of each monitoring cycle, the vehicle passage action in the first optimized monitoring image is identified, and a number of vehicle passage images corresponding to each vehicle passage action in the initial matching period are extracted to construct an image allocation reference data set; specifically, the method comprises: in the initial matching period of each monitoring cycle, a vehicle detection algorithm is used to identify the vehicle passage action in the first optimized monitoring image, and a passing vehicle anchor frame is generated in the first optimized monitoring image; based on the passing vehicle anchor frame, a number of vehicle passage images corresponding to each vehicle passage action in the initial matching period are extracted, and the number of vehicle passage images are used as the image allocation reference data set for each vehicle passage action; According to the image allocation reference data set, a matching relationship between the traffic camera and the image processing terminal is established; specifically comprising: identifying the target vehicle traffic image in the image allocation reference data set for each vehicle traffic action of each traffic camera within the initial matching period of each monitoring cycle, and generating the hardware optimization standard coefficient of each traffic camera; wherein the target vehicle traffic image is configured as an image whose vehicle RGB color value in the image allocation reference data set is abnormal compared with the vehicle RGB color value of the non-target vehicle traffic image in the image allocation reference data set, and the proportion of the target vehicle traffic image in the image allocation reference data set is less than the proportion of the non-target vehicle traffic image in the image allocation reference data set; wherein the hardware optimization standard coefficient includes a number of optimized pixel points with RGB color value adjustment values, and the optimized pixel points are The image processing unit is configured as a set of pixel points with abnormal RGB color values ​​of vehicles in all target vehicle images, and the RGB color value adjustment value is configured as the average value of the RGB color difference between the target vehicle traffic image and the non-target vehicle traffic image at the same position of the vehicle for each vehicle traffic action in the initial matching period; the number of target vehicle traffic images for all vehicle traffic actions in the initial matching period of each monitoring cycle is calculated as the image processing requirement of each traffic camera; based on the image processing requirement of each traffic camera and the image processing capability of the image processing terminal, the traffic camera is matched with the image processing terminal to establish a matching relationship between the traffic camera and the image processing terminal; wherein the image processing capability of the image processing terminal is configured as the number of standard test pictures processed by the image processing terminal per unit time; The first optimized monitoring image of the image processing period of each monitoring cycle is sent to an image processing terminal with a matching relationship, and a second color conversion is performed to obtain a second optimized monitoring image; specifically including: sending the first optimized monitoring image and the hardware optimization standard coefficient of each traffic camera in the image processing period of each monitoring cycle to an image processing terminal with a matching relationship; driving each image processing terminal to perform a second color conversion on the first optimized monitoring image according to the received hardware optimization standard coefficient to obtain a second optimized monitoring image.

2. The image processing method based on artificial intelligence as claimed in claim 1, It is characterized in that Based on the identification information of each traffic camera, the steps of querying the regional environment information of each traffic camera specifically include: Based on the identification information of each traffic camera, query the location information of each traffic camera in the traffic camera deployment data list; The location information is used to access the regional environment database to extract the current regional environment information of each traffic camera; wherein the regional environment information includes a plurality of regional environment data that affect the color recognition of the traffic monitoring image.

3. The image processing method based on artificial intelligence as claimed in claim 1, It is characterized in that Before the step of using the regional environment information as a regional environment feature set and inputting a pre-trained image color environment optimization model to obtain output image color environment optimization parameters for each traffic camera, the method further includes: Acquire test color images of standard color sample images collected by traffic cameras under different regional environmental feature combinations; wherein the regional environmental feature combination includes one or more of the following: solar altitude angle, atmospheric component content, cloud thickness parameter, and street light illumination color; A training sample set is constructed based on the standard color value of each standard color block in the standard color sample image, the combination of environmental features of different regions, and the photographed color value of each standard color block in the test color image; The training sample set is input into the constructed initial neural network model for training, and an image color environment optimization model is obtained after the training times reaches a preset number or the training converges.

4. An image processing device based on artificial intelligence, It is characterized in that include: An acquisition module, used to acquire traffic monitoring information of a target area; wherein the traffic monitoring information includes identification information of a plurality of traffic cameras and a traffic monitoring image captured by each traffic camera; A query module, used to query the regional environment information of each traffic camera based on the identification information of each traffic camera; A first execution module is used to use the regional environment information to perform a first color conversion on the traffic monitoring image taken by each traffic camera to obtain a first optimized monitoring image; specifically comprising: taking the regional environment information as a regional environment feature set, inputting a pre-trained image color environment optimization model, and obtaining output image color environment optimization parameters of each traffic camera; wherein the image color environment optimization parameters are configured as a numerical adjustment ratio of the RGB color value of each color channel; based on the image color environment optimization parameters, performing a first color conversion on the RGB color value of each pixel in the traffic monitoring image taken by each camera to obtain a first optimized monitoring image; wherein the first color conversion is configured to multiply the RGB color value of each pixel in the traffic monitoring image by the numerical adjustment ratio of the corresponding color channel, and using the multiplication result as the RGB color value of the corresponding pixel in the first optimized monitoring image; The recognition module is used to recognize the vehicle passage action in the first optimized monitoring image during the initial matching period of each monitoring cycle, extract a number of vehicle passage images corresponding to each vehicle passage action during the initial matching period, and construct an image allocation reference data set; specifically comprising: during the initial matching period of each monitoring cycle, using a vehicle detection algorithm to recognize the vehicle passage action in the first optimized monitoring image, and generating a passing vehicle anchor frame in the first optimized monitoring image; based on the passing vehicle anchor frame, extracting a number of vehicle passage images corresponding to each vehicle passage action during the initial matching period, and using the number of vehicle passage images as the image allocation reference data set for each vehicle passage action; An establishment module is used to establish a matching relationship between a traffic camera and an image processing terminal according to an image allocation reference data set; specifically comprising: identifying a target vehicle traffic image in an image allocation reference data set for each vehicle traffic action of each traffic camera within an initial matching period of each monitoring cycle, and generating a hardware optimization standard coefficient for each traffic camera; wherein the target vehicle traffic image is configured as an image whose vehicle RGB color value in the image allocation reference data set is abnormal compared to the vehicle RGB color value of a non-target vehicle traffic image in the image allocation reference data set, and the proportion of the target vehicle traffic image in the image allocation reference data set is less than the proportion of the non-target vehicle traffic image in the image allocation reference data set; wherein the hardware optimization standard coefficient includes a number of optimized pixel points with RGB color value adjustment values, and the optimized The pixel points are configured as a set of pixel points with abnormal vehicle RGB color values ​​in all target vehicle images, and the RGB color value adjustment value is configured as the average value of the RGB color difference between the target vehicle traffic image and the non-target vehicle traffic image at the same position of the vehicle with respect to each color channel in each vehicle traffic action in the initial matching period; the number of target vehicle traffic images for all vehicle traffic actions in the initial matching period of each monitoring cycle is calculated as the image processing requirement of each traffic camera; based on the image processing requirement of each traffic camera and the image processing capability of the image processing terminal, the traffic camera is matched with the image processing terminal to establish a matching relationship between the traffic camera and the image processing terminal; wherein the image processing capability of the image processing terminal is configured as the number of standard test pictures processed by the image processing terminal in a unit time; The second execution module is used to send the first optimized monitoring image of the image processing time period of each monitoring cycle to an image processing terminal with a matching relationship, perform a second color conversion, and obtain a second optimized monitoring image; specifically includes: sending the first optimized monitoring image and the hardware optimization standard coefficient of each traffic camera in the image processing time period of each monitoring cycle to the image processing terminal with a matching relationship; driving each image processing terminal to perform a second color conversion on the first optimized monitoring image according to the received hardware optimization standard coefficient to obtain a second optimized monitoring image.

5. An image processing device based on artificial intelligence, It is characterized in that The artificial intelligence-based image processing device includes: a memory, a processor, and an artificial intelligence-based image processing program stored in the memory and executable on the processor. When the artificial intelligence-based image processing program is executed by the processor, the steps of the artificial intelligence-based image processing method as described in any one of claims 1 to 3 are implemented.

6. An artificial intelligence-based storage medium, It is characterized in that The artificial intelligence-based storage medium stores an artificial intelligence-based image processing program, and when the artificial intelligence-based image processing program is executed by the processor, the steps of the artificial intelligence-based image processing method as described in any one of claims 1 to 3 are implemented.

Images