Accuracy improving method of automobile data recorder and related device

Abstract

The invention discloses an accuracy improving method of an automobile data recorder and a related device.The method comprises the steps that a first light intensity value in a target shooting area corresponding to a camera is obtained, and when the first light intensity value is smaller than a light intensity threshold value, a light supplementing signal is generated based on the first light intensity value, performing light supplement on the target shooting area based on the light supplement signal to obtain a second light intensity value in the target shooting area, when the second light intensity value is greater than a preset light intensity value, obtaining a first image corresponding to the target shooting area, and when the preset light intensity value is greater than a light intensity threshold value, determining a first quality evaluation value corresponding to the first image, and when the first quality evaluation value is smaller than a quality evaluation threshold value, performing noise reduction processing on the first image to obtain a second image, and finally storing the second image in the automobile data recorder. By adopting the embodiment of the invention, the image recording accuracy of the automobile data recorder is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of driving recorders, and particularly to a method for improving the accuracy of a driving recorder and related devices. Background Art

[0002] As an important vehicle-mounted device, a driving recorder plays a crucial role in modern transportation. It can record the images and related information during the vehicle driving in real time, providing key evidence support for the determination of liability in traffic accidents, daily driving safety monitoring, etc. However, in the actual use process, the current driving recorder faces many challenges, and the recording accuracy of the driving recorder is not high. Therefore, how to improve the accuracy of the recorded images of the driving recorder is an urgent problem to be solved. Summary of the Invention

[0003] The embodiments of the present application provide a method for improving the accuracy of a driving recorder and related devices, which improves the accuracy of the recorded images of the driving recorder.

[0004] In a first aspect, the embodiments of the present application provide a method for improving the accuracy of a driving recorder, which is applied to a driving recorder including a camera. The method includes:

[0005] Obtaining a first light intensity value in a target shooting area corresponding to the camera;

[0006] When the first light intensity value is less than a light intensity threshold, generating a fill light signal based on the first light intensity value;

[0007] Performing fill light on the target shooting area based on the fill light signal to obtain a second light intensity value in the target shooting area;

[0008] When the second light intensity value is greater than a preset light intensity value, obtaining a first image corresponding to the target shooting area; the preset light intensity value is greater than the light intensity threshold;

[0009] Determining a first quality evaluation value corresponding to the first image;

[0010] When the first quality evaluation value is less than a quality evaluation threshold, performing noise reduction processing on the first image to obtain a second image; the second quality evaluation value corresponding to the second image is greater than the quality evaluation threshold;

[0011] Storing the second image in the driving recorder.

[0012] Second aspect, an accuracy improvement device for a driving recorder provided by an embodiment of the present application is applied to a driving recorder, and the driving recorder includes a camera; the accuracy improvement device of the driving recorder includes: an acquisition unit and a processing unit;

[0013] The acquisition unit is configured to acquire a first light intensity value within a target shooting area corresponding to the camera;

[0014] The processing unit is configured to generate a fill light signal based on the first light intensity value when the first light intensity value is less than a light intensity threshold;

[0015] Perform fill light on the target shooting area based on the fill light signal to obtain a second light intensity value within the target shooting area;

[0016] When the second light intensity value is greater than a preset light intensity value, acquire a first image corresponding to the target shooting area; the preset light intensity value is greater than the light intensity threshold;

[0017] Determine a first quality evaluation value corresponding to the first image;

[0018] When the first quality evaluation value is less than a quality evaluation threshold, perform noise reduction processing on the first image to obtain a second image; the second quality evaluation value corresponding to the second image is greater than the quality evaluation threshold;

[0019] Store the second image in the driving recorder.

[0020] Third aspect, an embodiment of the present invention provides an electronic device, including: a processor, a memory, a communication interface, and one or more programs, wherein the one or more programs are stored in the memory and are configured to be executed by the processor so that the electronic device executes the method as in the first aspect.

[0021] Fourth aspect, an embodiment of the present invention provides a computer-readable storage medium, and the computer-readable storage medium stores a computer program, and the computer program is executed by a processor to implement the method as in the first aspect.

[0022] Fifth aspect, an embodiment of the present invention provides a computer program product, and the computer program product includes a non-transitory computer-readable storage medium storing a computer program, so that a computer executes the method as in the first aspect.

[0023] Implementing the embodiments of the present invention has the following beneficial effects:

[0024] It can be seen that the method for improving the accuracy of a driving recorder described in the embodiments of the present invention is applied to a driving recorder, which includes a camera. First, the first light intensity value in the target shooting area corresponding to the camera is obtained. When the first light intensity value is less than the light intensity threshold, a fill light signal is generated based on the first light intensity value. Then, the target shooting area is filled with light based on the fill light signal to obtain the second light intensity value in the target shooting area. When the second light intensity value is greater than the preset light intensity value, the first image corresponding to the target shooting area is then obtained. The preset light intensity value is greater than the light intensity threshold. The first quality evaluation value corresponding to the first image is determined. When the first quality evaluation value is less than the quality evaluation threshold, the first image is denoised to obtain a second image. The second quality evaluation value corresponding to the second image is greater than the quality evaluation threshold. Finally, the second image is stored in the driving recorder, thereby improving the accuracy of the images recorded by the driving recorder. Description of the Drawings

[0025] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the background art, the following will describe the drawings required for use in the embodiments of the present application or the background art.

[0026] Figure 1 It is a schematic structural diagram of a system for improving the accuracy of a driving recorder provided by an embodiment of the present application;

[0027] Figure 2 It is a flowchart of a method for improving the accuracy of a driving recorder provided by an embodiment of the present application;

[0028] Figure 3 It is a flowchart of determining the first quality evaluation value provided by an embodiment of the present application;

[0029] Figure 4 It is a flowchart of determining n reference noise intensity values provided by an embodiment of the present application;

[0030] Figure 5 It is a flowchart of determining the second image provided by an embodiment of the present application;

[0031] Figure 6 It is an example diagram of dividing historical images provided by an embodiment of the present application;

[0032] Figure 7 It is a schematic structural diagram of a device for improving the accuracy of a driving recorder provided by an embodiment of the present application;

[0033] Figure 8 It is a schematic structural diagram of an electronic device provided by an embodiment of the application. Detailed implementation manners

[0034] To enable those skilled in the art to better understand the solutions of this application, the technical solutions in the embodiments of this application will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of this application. Obviously, the described embodiments are only a part of the embodiments of this application, rather than all the embodiments. Based on the embodiments in this application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the scope of protection of this application.

[0035] The terms "first", "second", etc. in the description and claims of this application and the above-mentioned accompanying drawings are used to distinguish different objects, rather than to describe a specific order. In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device that includes a series of steps or units is not limited to the listed steps or units, but optionally further includes steps or units not listed, or optionally further includes other steps or units inherent to these processes, methods, products or devices.

[0036] Referring to " is a schematic structural diagram of an accuracy improvement system for a driving recorder provided by an embodiment of this application.

[0037] Please refer to Figure 1 , Figure 1 is a schematic structural diagram of an accuracy improvement system for a driving recorder provided by an embodiment of this application.

[0038] The accuracy improvement system 10 of the driving recorder includes a driving recorder 101 and a target vehicle 102, and the driving recorder 101 includes a camera.

[0039] There are various connection methods between the dash cam and the target vehicle. The specific connection path depends on the functional design of the dash cam, the vehicle configuration, and the usage scenario. The following are several common connection methods: Cigarette lighter connection: This is one of the most common connection methods. The dash cam draws power from the vehicle's cigarette lighter through the cigarette lighter interface. The cigarette lighter provides direct current of 12V or 24V (depending on the vehicle type). This connection method is simple and convenient, and both installation and disassembly are easy. It is suitable for most dash cams. For example, many dash cam products in the aftermarket are equipped with cigarette lighter plugs, and users only need to insert them into the vehicle's cigarette lighter to power the dash cam; Fuse box connection: Some dash cams choose to connect to the vehicle's fuse box to obtain power. This connection method requires certain professional knowledge because the appropriate fuse position needs to be correctly selected to ensure stable power supply for the dash cam while not affecting the normal operation of other electrical devices in the vehicle. Through the fuse box connection, functions such as the dash cam automatically turning on when the vehicle starts and delaying shutdown after the vehicle shuts off can be achieved, providing a more intelligent usage experience; The dash cam is built-in with a Wi-Fi module, and users can connect their mobile phones or other mobile devices to the dash cam's Wi-Fi hotspot. In this way, users can view the images captured by the dash cam in real time on the mobile device, perform setting adjustments, download videos, etc. For example, after parking, users can connect their mobile phones to the dash cam's Wi-Fi to check the situation around the vehicle to ensure vehicle safety; Bluetooth connection: It can achieve short-distance communication between the dash cam and devices such as mobile phones. Through Bluetooth connection, the dash cam can transmit video clips, device status information, etc. captured to the mobile phone, and at the same time, it can also receive control commands sent by the mobile phone. In addition, some dash cams also support connecting to Bluetooth headsets to achieve functions such as voice prompts, enhancing the usage convenience for users.

[0040] In this embodiment, when the target vehicle 102 is in motion, first, obtain the first light intensity value within the target shooting area corresponding to the camera. When the first light intensity value is less than the light intensity threshold, generate a fill light signal based on the first light intensity value, and then perform fill light on the target shooting area based on the fill light signal to obtain the second light intensity value within the target shooting area. When the second light intensity value is greater than the preset light intensity value, obtain the first image corresponding to the target shooting area, where the preset light intensity value is greater than the light intensity threshold. Then, determine the first quality evaluation value corresponding to the first image. When the first quality evaluation value is less than the quality evaluation threshold, perform noise reduction processing on the first image to obtain the second image, where the second quality evaluation value corresponding to the second image is greater than the quality evaluation threshold. Finally, store the second image in the dash cam 101.

[0041] It can be seen that by obtaining the first light intensity value of the target shooting area and performing supplementary lighting when the light is insufficient (less than the light intensity threshold), the shooting quality in low-light environments can be effectively improved. In scenes with dim light such as at night, underground parking lots, and tunnels, the supplementary lighting operation can make the light in the target area more sufficient, avoiding blurring and dimness of the captured images due to insufficient exposure, and ensuring clear presentation of key information such as road conditions, vehicle markings, and pedestrians. By setting a preset light intensity value and obtaining the image only when the second light intensity value is greater than this preset value after supplementary lighting, the light conditions during each shooting can be relatively stable. Stable light conditions can reduce the fluctuation of image quality caused by light changes, making the captured images consistent and facilitating subsequent analysis and viewing. By determining the first quality evaluation value of the first image and comparing it with the quality evaluation threshold, it is possible to accurately judge whether the image meets the usage requirements. This quantitative evaluation method avoids the uncertainty of judging image quality based solely on subjective feelings and provides an objective basis for subsequent processing decisions. If the image quality evaluation value is lower than the threshold, it indicates that there may be problems such as noise and blurring in the image, and further processing is required. Otherwise, it can be considered that the image quality is good. Through image quality evaluation, the dash cam can distinguish high-quality and low-quality images, perform targeted processing on low-quality images, and avoid unnecessary processing of all images, thereby optimizing the utilization of device resources and improving operation efficiency. When the first image quality evaluation value is less than the quality evaluation threshold, noise reduction processing is performed, which can effectively remove noise interference in the image and make the image clearer and smoother. Noise can cause problems such as graininess and blurring in the image, affecting the presentation of details and readability. After noise reduction processing, the target objects in the image are more clearly distinguishable, which helps to accurately identify important information such as license plate numbers, traffic signs, and pedestrian facial features. It enhances the accuracy and reliability of the dash cam's recording. In scenarios such as traffic accident liability determination and traffic violation recording, clear and noise-free images have higher credibility and effectiveness as evidence. The second image after noise reduction processing has higher quality and can better meet the requirements as evidence, providing strong support for the decision-making of relevant departments. Storing the processed second image in the dash cam can ensure the proper preservation of important information during the driving process. These images can be used as evidence when needed, such as for determining liability in the event of a traffic accident or restoring the scene in case of a dispute. The stored images are convenient for the vehicle owner or relevant personnel to view and analyze later. The vehicle owner can review the driving process at any time to understand information such as the vehicle's driving conditions and the surrounding environment. This not only improves the accuracy of the dash cam's recorded images but also enables the traffic management department to conduct traffic violation investigations and accident investigations by retrieving the stored images.

[0042] Please refer to Figure 2 , Figure 2 is a flowchart of a method for improving the accuracy of a dash cam provided by an embodiment of the present application, including but not limited to the following steps:

[0043] S201: Obtain the first light intensity value within the target shooting area corresponding to the camera.

[0044] In this embodiment, the light intensity, also known as the illumination intensity, represents the luminous flux of visible light received per unit area, and its unit is lux. In the scenario of the dashcam in this embodiment, the light intensity value within the target shooting area reflects the brightness of the ambient light in this area. This value is crucial for determining whether the current environment is suitable for taking clear images. For example, on a sunny day, the light intensity may reach several thousand or even tens of thousands of lux, while on a road without streetlights at night, the light intensity may be only a few lux or even lower.

[0045] The shooting effects of the dashcam vary greatly under different light conditions. If the light is too dim, the captured images will be blurred, dim, and many important details will be lost. If the light is too strong, it may cause overexposure of the images, which also affects the image quality. Therefore, by obtaining the first light intensity value within the target shooting area, it can provide a basis for subsequent supplementary lighting operations and image quality evaluation, ensuring that high-quality images are captured under suitable light conditions.

[0046] The dashcam integrates a light sensor internally. This sensor can sense the light intensity of the surrounding environment and convert it into an electrical signal. Common light sensors include photoresistors, photodiodes, etc. The resistance value of the photoresistor will change with the change of the light intensity. By measuring the change in its resistance value, the corresponding light intensity value can be obtained. The photodiode reflects the light intensity according to the magnitude of the current generated by light. The light sensor continuously senses the light situation in the target shooting area. After converting the optical signal into an electrical signal, it is converted into a digital signal through an analog-to-digital converter, and then transmitted to the main control chip of the dashcam for processing and analysis, and finally the first light intensity value is obtained.

[0047] In this embodiment, the image data captured by the camera itself is used to indirectly estimate the light intensity. In the image, the distribution of pixel values is related to the light intensity. Generally speaking, the higher the average pixel value of the whole image, the stronger the light. On the contrary, the lower the average pixel value, the weaker the light. The main control chip of the dashcam analyzes the initial image captured by the camera, calculates the average gray value of all pixels in the image (for a color image, it can be first converted into a gray image), and then, through the pre-established mapping relationship between the average gray value and the light intensity, converts the average gray value into the corresponding light intensity value.

[0048] It should be noted that if a built-in light sensor is used, its installation position and angle will affect the accuracy of the light intensity value. The sensor should be installed at a position where it can accurately sense the light in the target shooting area, avoiding being blocked or interfered by other light sources. In the actual environment, the light intensity may be affected by various factors, such as weather changes and reflections from surrounding buildings. Therefore, when obtaining the light intensity value, it is necessary to consider the influence of these factors and perform appropriate calibration and compensation to ensure that the obtained light intensity value is accurate and reliable. Since the ambient light changes continuously during vehicle driving, it is necessary to update the first light intensity value in real time so as to make decisions such as supplementary lighting in a timely manner and ensure the quality of the captured images.

[0049] S202: When the first light intensity value is less than the light intensity threshold, generate a supplementary lighting signal based on the first light intensity value.

[0050] In this embodiment, the light intensity has a crucial impact on the quality of the images captured by the dashcam. The light intensity threshold is a preset standard value, which represents the minimum light intensity required to ensure that the dashcam can capture clear and high-quality images. When the first light intensity value is less than the light intensity threshold, it indicates that the light conditions in the current target shooting area are poor. Direct shooting may result in dim, blurred images, losing important detail information and unable to meet the requirements of recording the driving process. Therefore, it is necessary to perform a supplementary lighting operation, and the supplementary lighting signal is an instruction to trigger the supplementary lighting device to work.

[0051] In the system of the dashcam, a suitable light intensity threshold will be preset according to a large amount of experimental data and actual usage requirements. This threshold is usually in lux. Different dashcams may set different light intensity thresholds according to the performance and design goals of their cameras. The main control chip of the dashcam will obtain the first light intensity value in real time and compare it with the preset light intensity threshold. If the first light intensity value is less than the light intensity threshold, it indicates that the current light conditions do not meet the shooting requirements and it is necessary to start the supplementary lighting operation. The generation of the supplementary lighting signal is closely related to the first light intensity value. Generally speaking, the larger the difference between the first light intensity value and the light intensity threshold, the more insufficient the light is and the stronger the supplementary lighting intensity is required. The main control chip of the dashcam will calculate appropriate supplementary lighting parameters, such as the brightness and duration of the supplementary lighting device, according to this difference, and then encode these parameters into a supplementary lighting signal. The supplementary lighting signal can be a digital signal and is transmitted to the supplementary lighting device through a specific communication protocol, or it can be an analog signal to directly control the working state of the supplementary lighting device.

[0052] It can be seen that in an environment with insufficient light, such as at night, in an underground parking lot or a tunnel, etc., the first light intensity value is often lower than the light intensity threshold. At this time, generating a fill light signal for fill light can effectively improve the light brightness of the target shooting area. Sufficient light can make the images captured by the driving recorder brighter. Objects that were originally blurred due to dim light, such as road signs, vehicle license plates, pedestrian faces, etc., can become clearly distinguishable. This helps to record more accurate and detailed driving information and provides strong evidence support for possible traffic accident liability determination, traffic violation records, etc. Images captured under low light conditions are prone to noise, making the images appear grainy, affecting the visual effect and information extraction. Fill light can improve the signal-to-noise ratio of the images, reduce the interference of noise, make the images smoother and more realistic. At the same time, appropriate fill light can also avoid the problem of color distortion of the images caused by insufficient light, making the captured images closer to the colors of the actual scene, further improving the quality and reliability of the images. Generating a fill light signal based on the first light intensity value realizes the automatic response and adjustment of the driving recorder to different lighting environments. Regardless of the lighting conditions encountered during the vehicle's driving process, the driving recorder can automatically determine whether fill light is needed according to the real-time light intensity situation and generate the corresponding fill light signal in a timely manner. This intelligent function enables the driving recorder to work stably in various complex lighting environments and provides continuous and reliable driving record services for users. Since the fill light signal is generated based on the first light intensity value, the driving recorder can dynamically adjust the intensity of the fill light according to the degree of insufficient light. When the difference between the first light intensity value and the light intensity threshold is large, it indicates that the light is severely insufficient, and the fill light signal will instruct the fill light device to provide stronger light; while when the difference is small, the fill light intensity will also be reduced accordingly. This dynamic adjustment method can not only ensure sufficient fill light effect in case of insufficient light, but also avoid energy waste and image overexposure problems caused by excessive fill light.

[0053] S203: Perform fill light on the target shooting area based on the fill light signal to obtain the second light intensity value within the target shooting area.

[0054] In this embodiment, a common supplementary lighting device for a driving recorder is an LED supplementary light. The LED supplementary light has the advantages of low energy consumption, high brightness, fast response speed, etc., and is very suitable for supplementary lighting of a driving recorder. After receiving the supplementary lighting signal, the supplementary lighting device will work according to the supplementary lighting parameters included in the signal. For example, if the supplementary lighting signal indicates that supplementary lighting needs to be continuously provided at a certain brightness for a period of time, the supplementary lighting device will adjust its output power to make the LED light reach the corresponding brightness and keep it on for the specified time. During the supplementary lighting process, the supplementary lighting device will evenly illuminate the target shooting area and increase the light intensity in this area. After the supplementary lighting operation has been carried out for a period of time, the driving recorder will obtain the light intensity value in the target shooting area again, that is, the second light intensity value, through an internal light sensor (such as a photoresistor, a photodiode, etc.) or a method of auxiliary analysis using the camera image data. By comparing the second light intensity value with the preset light intensity value, the effect of the supplementary lighting operation can be evaluated. If the second light intensity value is greater than the preset light intensity value, it means that the supplementary lighting operation has achieved the expected effect, and at this time, image shooting can be carried out; if the second light intensity value is still less than the preset light intensity value, it may be necessary to further adjust the supplementary lighting parameters and perform the supplementary lighting operation again until the light intensity meets the shooting requirements.

[0055] It can be seen that obtaining the second light intensity value can intuitively understand whether the supplementary lighting operation has achieved the expected effect. By comparing the second light intensity value with the preset light intensity value, if the second light intensity value is greater than the preset light intensity value, it means that the supplementary lighting is sufficient and meets the shooting requirements. If it is still less than the preset value, the supplementary lighting parameters can be further adjusted according to the difference, such as increasing the supplementary lighting brightness or extending the supplementary lighting time, so as to achieve more accurate supplementary lighting control. During the driving process of the vehicle, the ambient light will change continuously. By obtaining the second light intensity value in real time, the driving recorder can dynamically adjust the supplementary lighting strategy according to the change of light. For example, when the vehicle exits the tunnel and enters a bright environment, the second light intensity value will increase significantly. At this time, the supplementary lighting intensity can be automatically reduced or the supplementary lighting can be stopped to avoid energy waste and image overexposure caused by excessive supplementary lighting. Supplementary lighting is not only beneficial to the images captured by the driving recorder, but also can improve the driver's field of vision in low-light environments to a certain extent. The supplementary lighting device illuminates the road ahead and the surrounding environment of the vehicle, enabling the driver to observe potential dangers, such as obstacles, pedestrians or other vehicles, more clearly and react in advance, reducing the probability of traffic accidents.

[0056] S204: When the second light intensity value is greater than the preset light intensity value, obtain the first image corresponding to the target shooting area.

[0057] In this embodiment, the preset light intensity value is greater than the light intensity threshold. The preset light intensity value is a standard value determined through a large number of experiments and actual tests. It represents the minimum light intensity required for the dash cam to capture high-quality images. This value is greater than the light intensity threshold. The light intensity threshold is only used as an initial reference for determining whether supplementary lighting is needed, while the preset light intensity value is a key indicator to ensure that the captured images have sufficient brightness, clarity, and color reproduction. Different shooting scenarios and usage purposes have different requirements for image quality, and the preset light intensity value can be adjusted according to these needs. For example, in scenarios where it is necessary to clearly record license plate numbers or road signs, the preset light intensity value will be set relatively high to ensure that the captured images can accurately identify this information. In some scenarios where less detail is required, the preset light intensity value can be appropriately reduced.

[0058] When the first light intensity value is less than the light intensity threshold, the dash cam will generate a supplementary lighting signal based on this and perform supplementary lighting on the target shooting area to obtain the second light intensity value. Waiting for the second light intensity value to be greater than the preset light intensity value is to ensure that the supplementary lighting operation has achieved the expected effect and the light conditions in the target shooting area meet the requirements for shooting high-quality images. If shooting is carried out when the light intensity has not reached the standard, the images may still have problems such as insufficient brightness and blurriness. Real-time monitoring of the second light intensity value and shooting only when it is greater than the preset light intensity value can avoid poor image quality caused by premature shooting and prevent missing important driving images due to late shooting. This precise control mechanism ensures that the dash cam can obtain images under the best light conditions. When the second light intensity value is greater than the preset light intensity value, the camera of the dash cam will immediately activate the shooting function to obtain the first image corresponding to the target shooting area. These images can record various information during vehicle driving, such as road conditions, traffic signs, the situations of other vehicles and pedestrians, etc., providing important evidence for traffic accident liability determination, daily driving safety monitoring, etc. As the data collected by the dash cam, the first images can also be used for subsequent analysis and processing. For example, through image processing and computer vision algorithm analysis of the images, functions such as vehicle recognition, pedestrian detection, and lane departure warning can be realized, improving the safety and intelligence level of driving.

[0059] It can be seen that by first performing supplementary lighting and then shooting images under suitable light conditions, not only can the quality and reliability of the images captured by the dash cam be significantly improved, but also the high-quality images can more accurately reflect the actual driving situation, reducing information loss or misjudgment caused by light problems. For users, clear and accurate driving record images can provide a better user experience. Whether used as evidence in case of a traffic accident or viewed during daily review of the driving process, high-quality images can make users feel more at ease and satisfied.

[0060] S205: Determine the first quality assessment value corresponding to the first image.

[0061] In this embodiment, refer to Figure 3 , Figure 3 which is a flowchart for determining the first quality assessment value provided by the embodiment of the present application, including but not limited to the following steps:

[0062] S301: Determine the clarity and noise intensity values corresponding to the first image.

[0063] In this embodiment, exemplarily, determine the gradient magnitude corresponding to the first image. Specifically, the gradient of an image is used to describe the change rate of image pixel values, which reflects the severity of gray-scale changes in the image. In the image, the pixel values in the edge and detail regions change greatly, and the gradient values are also large, while in the smooth region, the pixel values change little, and the gradient values are also small. Therefore, by calculating the gradient magnitude of the image, the edge and detail information in the image can be captured, and then used to evaluate the clarity of the image. First, calculate the gradient components of the image in the horizontal and vertical directions, and then calculate the gradient magnitude of each pixel according to the gradient components of the image in the horizontal and vertical directions. Finally, a gradient magnitude image with the same size as the original image is obtained, where the value of each pixel represents the gradient magnitude at that position.

[0064] Exemplarily, obtain the first mapping relationship between the gradient magnitude and clarity. Specifically, there is a certain correlation between the gradient magnitude and image clarity. Generally speaking, the larger the gradient magnitude, the more obvious the edges and details in the image, and the clearer the image. Through a large number of experiments and data analysis, the mapping relationship between the gradient magnitude and clarity can be established. This mapping relationship can be a mathematical function, a look-up table, or a machine learning model. Collect a large number of image samples with different clarities, calculate the gradient magnitude and the corresponding clarity assessment value of each image (which can be obtained through subjective human evaluation or other objective methods). Then, analyze the relationship between the gradient magnitude and the clarity assessment value, and fit a suitable mathematical function, such as a linear function, a polynomial function, etc. Use machine learning algorithms, such as regression analysis, neural networks, etc., with the gradient magnitude as the input feature and the clarity assessment value as the output label to train the model. After training, the model can predict the corresponding clarity according to the input gradient magnitude.

[0065] Exemplarily, determine the clarity corresponding to the gradient magnitude based on the first mapping relationship to obtain the clarity corresponding to the first image. Specifically, substitute the calculated gradient magnitude of the first image into the first mapping relationship and calculate according to the specific form of the mapping relationship to obtain the corresponding clarity assessment value, that is, the clarity corresponding to the first image.

[0066] Exemplarily, the first image is divided into n image blocks, where n is an integer greater than 1, and the sizes of the n image blocks are the same. Specifically, dividing the image into multiple image blocks of the same size is to more carefully analyze the noise distribution of the image. Different image regions may have different noise characteristics. By analyzing each image block separately, the noise intensity of the entire image can be estimated more accurately.

[0067] Exemplarily, determine the reference noise intensity value corresponding to each of the n image blocks, and obtain n reference noise intensity values. Please refer to Figure 4 , Figure 4 which is a flowchart for determining n reference noise intensity values provided by an embodiment of the present application, including but not limited to the following steps:

[0068] S401: Obtain multiple pixel values within the first image block.

[0069] In this embodiment, the first image block is any one of the n image blocks. An image is composed of pixel points, and each pixel point has its corresponding pixel value. For a grayscale image, the pixel value is usually an integer between 0 and 255, representing the brightness of the pixel point. For a color image, the pixel value usually consists of the values of the red, green, and blue channels. Obtaining multiple pixel values within the first image block means extracting the pixel values corresponding to all pixel points within this specific image block for subsequent calculations. For an already read and segmented image block, the pixel values can be obtained using array indexing, so the multiple pixel values within the first image block can also be obtained using array indexing.

[0070] It can be seen that the noise distribution in an image is often uneven, and the noise intensity may vary in different regions. Dividing the image into multiple image blocks and obtaining the pixel values within the first image block can focus on the local region of the image and more accurately reflect the fluctuations of the pixel values in that region. Since noise causes random changes in pixel values, by analyzing these pixel values, the noise characteristics of the local region can be captured, thereby more accurately evaluating the noise intensity of the image block. For example, in an image captured by a dash cam, due to light refraction or local sensor failures, the noise in some regions may be more obvious than in others. By obtaining the pixel values of the image block, these local noise anomalies can be detected. By analyzing the pixel values of each image block, the reference noise intensity values of each image block are obtained, and then the noise level of the entire image can be comprehensively evaluated. This evaluation method from local to global is more detailed and accurate than directly analyzing the entire image. For example, when calculating the noise intensity of the entire image, a weighted average method can be used, and different weights are assigned according to the importance of each image block or the noise distribution, so as to obtain a result that is more in line with the actual situation. Different image blocks may have different noise characteristics. After obtaining the pixel values within the first image block, appropriate noise reduction algorithms and parameters can be selected according to the characteristics of these pixel values. For image blocks with higher noise intensity, more powerful noise reduction algorithms can be used, while for image blocks with lower noise intensity, relatively mild noise reduction algorithms can be used to avoid losing image details due to excessive noise reduction. For example, when processing an image containing complex textures and noise, processing according to the pixel value characteristics of different image blocks can effectively remove noise while better retaining the detail information of the image.

[0071] S402: Determine the pixel value variance corresponding to the multiple pixel values.

[0072] In this embodiment, variance is a statistic that measures the degree of dispersion of a set of data. In an image, noise causes random fluctuations in pixel values, resulting in pixel values deviating from their normal distribution. Therefore, by calculating the variance of the pixel values within an image block, the degree of dispersion of the pixel values within the image block can be reflected, and then the noise intensity can be inferred. If the pixel value variance is large, it indicates that the pixel value fluctuations are large, and the noise in the image block may be relatively severe. On the contrary, if the variance is small, it indicates that the pixel values are relatively stable, and the noise may be less.

[0073] It can be seen that image noise will cause random fluctuations in pixel values, causing pixel values ​​to deviate from their normal distribution. Variance is a statistic that measures the degree of data dispersion. By calculating the variance of pixel values, the degree of fluctuation of pixel values ​​in an image block can be quantified, which in turn reflects the noise intensity of the area. The larger the variance, the more drastic the fluctuation of pixel values ​​and the more serious the noise in the image block. Conversely, the smaller the variance, the lower the noise intensity. For example, in an image taken by a dashcam, if the variance of the pixel values ​​of an image block is large, it may mean that the area is subject to more noise interference, such as sensor noise, electromagnetic interference, etc. The details in the image will also cause changes in pixel values, but this change is different from the random fluctuations caused by noise. By calculating the variance, noise and image details can be distinguished to a certain extent. Generally speaking, the changes in pixel values ​​caused by noise are random, while the changes in image details have certain regularity. By analyzing the size and distribution of the variance, it can be determined whether the image block mainly contains noise or actual image details, thereby more accurately evaluating the image quality.

[0074] S403: Obtain a second mapping relationship between pixel value variance and reference noise intensity value.

[0075] In this embodiment, there is a certain correlation between the pixel value variance and the reference noise intensity value, but they are not simply equal. The second mapping relationship is a way to describe this correlation. It can be a mathematical function, a lookup table, or a model obtained through machine learning training. Through this mapping relationship, the calculated pixel value variance can be converted into the corresponding reference noise intensity value. First, a large number of image samples with different noise levels are collected, and the pixel value variance and the corresponding reference noise intensity value of each image block are calculated (which can be obtained by professional noise measurement equipment or other standard methods). Then, the relationship between these data is analyzed to fit a suitable mathematical function. Machine learning algorithms such as linear regression and decision trees can also be used to train the model with pixel value variance as input features and reference noise intensity values ​​as output labels. After the training is completed, the model can predict the corresponding reference noise intensity value based on the input pixel value variance.

[0076] S404: Determine a reference noise intensity value corresponding to the pixel value variance based on the second mapping relationship, and obtain a reference noise intensity value corresponding to the first image block.

[0077] In this implementation, the pixel value variance of the first image block calculated above is substituted into the second mapping relationship, and calculation is performed according to the specific form of the mapping relationship to obtain the corresponding reference noise intensity value.

[0078] After determining the reference noise intensity value corresponding to the first image block, since the first image block is any one of the n image blocks, and the determination method of the reference noise intensity value corresponding to each image block in the n image blocks is the same as that of the reference noise intensity value corresponding to the first image block, the reference noise intensity value corresponding to each image block in the n image blocks can be determined according to the determination method of the reference noise intensity value corresponding to the first image block, and n reference noise intensity values are obtained.

[0079] It can be seen that the noise distribution in different regions of the image is often uneven. Dividing the image into multiple image blocks and calculating the reference noise intensity value for each image block separately can capture the noise characteristics of local regions in the image in detail. After obtaining the reference noise intensity value of each image block, combining these values can more accurately evaluate the noise level of the entire image. Compared with directly evaluating the noise of the entire image, this local-to-global evaluation method takes into account the unevenness of the image noise distribution and avoids evaluation errors caused by the noise in some regions masking the noise characteristics of other regions, thus more truly reflecting the actual noise situation of the image. The noise intensities of different image blocks are different, and the corresponding noise reduction requirements are also different. According to the reference noise intensity value of each image block, the most suitable noise reduction algorithm and parameters can be selected for it. For image blocks with high noise intensity, a stronger noise reduction algorithm is used, and for image blocks with low noise intensity, a mild noise reduction algorithm is used to avoid losing image details due to excessive noise reduction. When extracting image features, noise will interfere with the accurate extraction of features. An accurate reference noise intensity value helps to better suppress the influence of noise during the feature extraction process and improve the accuracy of feature extraction. In image matching and recognition tasks, noise will reduce the accuracy of matching and the success rate of recognition. By evaluating and processing the noise of each image block, the quality of the image can be improved and the ability of image matching and recognition can be enhanced.

[0080] Exemplarily, the noise intensity value corresponding to the first image is determined based on the n noise intensity values. Specifically, the average value of the n noise intensity values can be calculated and used as the noise intensity value of the first image. It can also be considered that the contributions of different image blocks to the overall noise may be different, and different weights can be assigned to the reference noise intensity values of each image block, and then weighted averaging is performed. For example, the weights can be determined according to factors such as the position and content of the image block.

[0081] It can be seen that by calculating the gradient magnitude of the image and combining it with the pre-established first mapping relationship to determine the clarity, the clarity of the image can be quantified. Compared with subjective judgment, this quantified evaluation method is more objective and accurate. In the actual application of a dash cam, clear images are crucial for accurately recording road conditions, vehicle information, pedestrian features, etc. The gradient magnitude can sensitively capture the changes in pixel values in the image, and the edge and detail regions of the image are precisely where the pixel values change significantly. Therefore, the clarity determined based on the gradient magnitude can well reflect the richness of details in the image. For the images captured by the dash cam, accurately grasping the detailed features of the image helps to identify important information such as license plate numbers and traffic signs. The image is divided into multiple image blocks of the same size, the reference noise intensity value of each image block is calculated separately, and then the noise intensity value of the entire image is obtained comprehensively. This method takes into account the noise conditions of both the local and the whole image. Different image regions may be affected by different degrees of noise interference. By analyzing each image block separately, the distribution characteristics of the image noise can be understood more precisely. The accurate noise intensity value can provide an important basis for subsequent noise reduction processing. According to the noise intensity of the image, selecting appropriate noise reduction algorithms and parameters can effectively remove noise while retaining the detail information of the image to the greatest extent. Evaluating by combining the clarity and the noise intensity value of the image can comprehensively measure the quality of the image from multiple dimensions. Clarity reflects the sharpness and detail performance of the image, while the noise intensity reflects the purity of the image. By considering these two factors comprehensively, it can be more accurately judged whether the image meets the requirements of actual applications. For a dash cam, high-quality image recording is its core function. By accurately evaluating the clarity and noise intensity of the image, the dash cam can automatically adjust the shooting parameters or perform image processing according to the actual situation to improve the image quality.

[0082] S302: Determine a first reference quality evaluation value corresponding to the clarity and a second reference quality evaluation value corresponding to the noise intensity value.

[0083] In this embodiment, the first reference quality evaluation value is a numerical value obtained by quantitatively evaluating the characteristic of image clarity and is used to measure the contribution of image clarity to the overall image quality. A clear image enables the observer to more clearly identify the objects, details, and features in the image. The second reference quality evaluation value is the result of quantitatively evaluating the characteristic of image noise intensity and reflects the degree of influence of noise on the image quality. Noise will interfere with the normal display of the image, making the image blurred, grainy, and reducing the visual effect and interpretability of the image.

[0084] It can be a mapping relationship between a preset clarity and a reference quality evaluation value, an empirical relationship established between clarity and quality evaluation value through a large number of experiments and actual observations, or a relationship obtained by using machine learning algorithms such as support vector machines and neural networks to train a large amount of image data. Based on this mapping relationship, the first reference quality evaluation value corresponding to the clarity can be determined. It can be a mapping relationship between a preset noise intensity value and a reference quality evaluation value, a relationship determined by statistical analysis of images with different noise intensities to determine the relationship between noise intensity and quality evaluation value, or a relationship obtained by simulating different degrees of noise added to clean images and then having observers evaluate the quality of these noisy images. Based on this mapping relationship, the second reference quality evaluation value corresponding to the noise intensity value can be determined.

[0085] It can be seen that clarity and noise intensity are two important factors affecting image quality. By determining the first reference quality evaluation value and the second reference quality evaluation value, these two factors can be quantified, providing a basis for subsequent comprehensive evaluation of the overall image quality. These two reference quality evaluation values can provide important decision-making basis for image processing. If the first reference quality evaluation value is low, it indicates that the image clarity is insufficient and enhancement measures such as sharpening may be needed. If the second reference quality evaluation value is low, it indicates that the image has a large amount of noise and noise reduction processing is required. Based on these evaluation values, appropriate image processing algorithms and parameters can be selected targeted to improve the image quality.

[0086] S303: Determine the first quality evaluation value corresponding to the first image based on the first reference quality evaluation value and the second reference quality evaluation value.

[0087] In this embodiment, exemplarily, obtain the first reference weight corresponding to the first reference quality evaluation value and the second reference weight corresponding to the second reference quality evaluation value, where the sum of the first reference weight and the second reference weight is 1. Specifically, the first reference weight and the second reference weight respectively represent the proportions of image clarity and noise intensity in the overall image quality evaluation. Since these two factors jointly affect image quality and their combined effect constitutes the final image quality evaluation, the sum of their weights is 1.

[0088] Exemplarily, obtain the resolution of the first image. Specifically, image resolution refers to the number of pixels contained in each unit length of the image, usually represented by multiplying the number of horizontal pixels by the number of vertical pixels. Resolution reflects the fineness of the image. A high-resolution image contains more detailed information. A higher resolution means that the image contains more pixels, and these additional pixels can record more image details. When we view a high-resolution image, since there are enough pixels to depict features such as the outline and texture of an object, we can see various details in the image more clearly. For example, in a high-resolution landscape photo, we can clearly see the veins of the leaves, the texture of the distant mountains, etc. On the contrary, due to the limited number of pixels, a low-resolution image will lose a lot of details when depicting an object, making the image look blurry and difficult to distinguish the fine features of the object. When performing a magnification operation on the image, the advantage of a high-resolution image is more obvious because it itself has rich pixel information. After being magnified to a certain extent, it can still maintain a relatively clear image quality, and the edges and details of the object can still be presented relatively clearly.

[0089] Exemplarily, determine a first optimization factor corresponding to the resolution. Specifically, it can be a mapping relationship between a preset resolution and an optimization factor. Based on this mapping relationship, the first optimization factor corresponding to the resolution can be determined. The first optimization factor is a coefficient related to the image resolution and is used to adjust the first reference weight and the second reference weight according to the image resolution. Different resolutions may require different weight allocations to more accurately evaluate the image quality. For example, a high-resolution image may itself contain more details, and the importance of clarity is relatively higher. In this case, the first optimization factor may cause the first reference weight to increase; while a low-resolution image may be more affected by noise, and the weights need to be adjusted to highlight the influence of the noise intensity.

[0090] Exemplarily, optimize the first reference weight according to the first optimization factor to obtain a first target weight. Specifically, the first target weight is calculated according to the following formula:

[0091] First target weight = First reference weight × (1 + First optimization factor);

[0092] According to the above formula, the first reference weight can be optimized according to the first optimization factor to obtain the first target weight.

[0093] Exemplarily, adjust the second reference weight based on the first target weight to obtain a second target weight, where the sum of the first target weight and the second target weight is 1. Specifically, since the sum of the first target weight and the second target weight must be 1, the second target weight can be obtained by subtracting the first target weight from 1.

[0094] Exemplarily, the first quality evaluation value is calculated based on the first reference quality evaluation value, the second reference quality evaluation value, the first target weight, and the second target weight. Specifically, the first target weight is calculated according to the following formula:

[0095] First quality evaluation value = First reference quality evaluation value × First target weight + Second reference quality evaluation value × Second target weight;

[0096] According to the above formula, the first quality evaluation value can be calculated based on the first reference quality evaluation value, the second reference quality evaluation value, the first target weight, and the second target weight.

[0097] It can be seen that by optimizing the first reference weight according to the first optimization factor to obtain the first target weight and correspondingly adjusting the second reference weight to obtain the second target weight, the influence degree of clarity and noise intensity on image quality at different resolutions can be more accurately reflected. For example, for high-resolution images, the first target weight is appropriately increased to highlight the importance of clarity, and for low-resolution images, the second target weight is increased to emphasize the influence of noise intensity. This can avoid unreasonable weight allocation for clarity and noise intensity when evaluating image quality, thereby improving the accuracy of the evaluation result. An accurate image quality evaluation result can provide clear guidance for image processing. According to the first quality evaluation value, problems in terms of clarity and noise intensity of the image can be judged, and then appropriate image processing algorithms and parameters can be selected accordingly. For example, if the evaluation result shows that the image clarity is low, a sharpening algorithm can be used for enhancement, and if the noise intensity is high, noise reduction processing is required. Through the evaluation result after optimizing the weights, the key points and degrees of processing can be determined more accurately, improving the effect of image processing. By quantitatively evaluating the quality of different images, images that meet the requirements can be quickly screened out, improving work efficiency and the accuracy of decision-making.

[0098] S206: When the first quality evaluation value is less than the quality evaluation threshold, perform noise reduction processing on the first image to obtain a second image.

[0099] In this embodiment, the second quality evaluation value corresponding to the second image is greater than the quality evaluation threshold. When the first quality evaluation value is less than the quality evaluation threshold, it indicates that the quality of the first image does not meet the expected standard, and there may be problems such as more noise and insufficient clarity, which cannot well meet the subsequent usage requirements of the driving recorder. Therefore, it is necessary to perform noise reduction processing on the first image. Please refer to Figure 5 , Figure 5 which is a flowchart for determining a second image provided by an embodiment of the present application, including but not limited to the following steps:

[0100] S501: Obtain the historical recorded images in the driving recorder.

[0101] In this embodiment, during the use of the driving recorder, images during the vehicle driving process are continuously recorded. These historical recorded images contain rich scene information. The purpose of obtaining these historical recorded images is to provide reference data for subsequent noise reduction processing. Since images of the same shooting scene at different times may have similarities, these similarities can be used to effectively remove the noise in the current image. Through the storage system of the driving recorder, such as a memory card or an internal hard disk, the historical image data stored therein is read. In actual implementation, the built-in software interface of the driving recorder or relevant file reading functions can be used to obtain these images.

[0102] S502: Perform noise reduction processing on the n image blocks based on the historical recorded images to obtain n target image blocks.

[0103] In this embodiment, exemplarily, determine m similar image blocks corresponding to the first image block in the historical recorded images. Exemplarily, divide the historical recorded images into k candidate image blocks. The k candidate image blocks are all of the same size as the first image block, and k is an integer greater than m. Please refer to Figure 6 , Figure 6 which is an exemplary diagram of dividing historical images provided by the embodiment of the present application. The historical image 60 is divided into multiple candidate image blocks of equal size.

[0104] Exemplarily, determine the pixel difference between each image block in the k candidate image blocks and the first image block to obtain k pixel differences. Specifically, for each candidate image block and the first image block, calculate the pixel value difference pixel by pixel, and then summarize these differences (such as summing, summing the squares, etc.) to obtain a value representing the degree of difference between these two image blocks, that is, the pixel difference. After calculation, k pixel differences will be obtained, and each difference corresponds to the similarity between a candidate image block and the first image block.

[0105] Exemplarily, determine m pixel differences less than a preset pixel difference among the k pixel differences. Specifically, the preset pixel difference is a preset threshold for determining whether two image blocks are similar enough. Only when the pixel difference is less than this threshold, the corresponding candidate image block is considered similar to the first image block. Compare the obtained k pixel differences one by one to find the pixel differences less than the preset pixel difference. Since m similar image blocks are required subsequently, when screening, it may be necessary to sort the pixel differences that meet the conditions and select the smallest m pixel differences.

[0106] Exemplarily, candidate image blocks corresponding to the m pixel differences are determined to obtain the m similar image blocks. Specifically, each pixel difference corresponds to a candidate image block because the pixel difference is calculated by this candidate image block and the first image block. According to the selected m pixel differences, the candidate image blocks corresponding to them are found, and these candidate image blocks are the image blocks similar to the first image block. Subsequently, information of these similar image blocks, such as the average value of pixel values, can be used to process the first image block, for example, noise reduction processing, to improve the quality of the first image block.

[0107] It can be seen that the similar image blocks can provide more context information and feature samples for the first image block. When performing image feature extraction and analysis, by combining the features of multiple similar image blocks, the features of the first image block can be described more comprehensively and accurately, which helps to improve the accuracy of tasks such as image recognition and classification. A large amount of information is contained in the historical record images. By dividing them into candidate image blocks and comparing them with the first image block, the potential value in these historical data can be mined. Even though the historical record images and the current image are taken at different times, the information of similar scenes can provide important references for the processing of the current image. Using the similar image blocks in the historical record images for processing avoids starting complex analysis and processing from scratch for each new image. By reusing the information in the historical data, computing resources and time costs can be saved to a certain extent, and the processing efficiency of the system can be improved. In image matching and retrieval tasks, the determination of similar image blocks can be used as a priori knowledge to help find images similar to the first image block more accurately.

[0108] Exemplarily, pixel values corresponding to each of the m similar image blocks are determined to obtain m pixel values, where m is an integer greater than 1. Specifically, for the determined m similar image blocks, their pixel values are extracted respectively. The pixel value is the numerical value of each pixel point in the image. For a grayscale image, the pixel value is usually an integer between 0 and 255. For a color image, the pixel value usually consists of the values of the red, green, and blue channels. In actual operation, the pixel value of each pixel point can be obtained through the pixel matrix of the image.

[0109] Exemplarily, the average pixel value corresponding to the m pixel values is determined. Specifically, the pixel values of the m similar image blocks are averaged to obtain the average pixel value. For a grayscale image, the average value of the m pixel values is directly calculated. For a color image, the pixel values of the red, green, and blue channels are averaged respectively. The calculation of the average pixel value can effectively reduce the influence of noise because noise is usually randomly distributed, and the influence of noise can be offset by the averaging operation.

[0110] Exemplarily, update the pixel values of the first image block to the average pixel value to obtain the target image block corresponding to the first image block. Specifically, replace the original pixel values of the first image block with the calculated average pixel value, thereby obtaining the target image block after noise reduction processing. This method is based on the assumption that the pixel values in similar image blocks can better represent the real scene information, while noise is random. By replacing the original pixel values with the average pixel value, the noise in the first image block can be removed, making the image clearer.

[0111] According to the determination method of the target image block corresponding to the first image block, the n image blocks can be denoised based on the historical record image to obtain n target image blocks.

[0112] S503: Determine the second image based on the n target image blocks.

[0113] In this embodiment, the n target image blocks obtained after noise reduction processing are recombined according to the original positional relationship to obtain the second image after noise reduction. This process is actually applying the processing result of the image block to the entire image, thereby realizing the noise reduction processing of the entire image. Through this noise reduction method based on the historical record image, the information of historical data can be fully utilized to effectively remove the noise in the image and improve the quality of the image.

[0114] It can be seen that in actual use, the images captured by the dashcam are easily affected by various factors such as light and sensor performance, and noise is easily generated. Noise will make the image blurred and grainy, seriously affecting the clarity and recognizability of the image. By finding similar image blocks in the historical record image and calculating the average pixel value, the noise in the current image block can be effectively reduced. Because noise is usually randomly distributed, when taking the average value in multiple similar image blocks, the random noise will cancel each other out, making the target image block smoother and clearer, improving the overall quality of the image. Using the similar image blocks in the historical record image for processing is based on the similarity of image content, which can retain the real details of the image to the greatest extent while removing noise, making the image more real and natural. By effectively using the historical record image, the utilization rate of data is improved, and data waste is reduced. At the same time, the high-quality image after noise reduction can better meet the needs of users, such as accident liability determination and dashcam record viewing, thereby improving the accuracy and reliability of the dashcam.

[0115] S207: Store the second image in the dashcam.

[0116] In this embodiment, one of the core functions of the driving recorder is to record various scenarios during vehicle driving. As a high-quality image after optimization processing, the second image can record information such as the surrounding environment, road conditions, and traffic signs during vehicle driving more clearly and accurately. These image data can provide a complete driving record for the vehicle owner, facilitating them to review the journey and understand various situations during driving. Over time, the driving recorder will store a large number of second images, forming an image data set in a time series. These data can reflect the driving conditions of the vehicle at different times and locations, which is of great value for analyzing traffic flow, road usage, etc. At the same time, it can also provide data support for the long-term use and maintenance of the vehicle. For example, by analyzing the images at different time periods, the impact of changes in the vehicle's surrounding environment on the vehicle can be understood. The driving recorder usually has a certain storage capacity and needs to manage the stored second images reasonably. Generally, the driving recorder will adopt a cyclic recording method. When the storage capacity is full, the earliest image data will be automatically overwritten, which can ensure that the recorder always saves the latest driving record and also avoids data loss due to insufficient storage capacity.

[0117] In summary, implementing the embodiment of the present invention has the following beneficial effects:

[0118] It can be seen that the method for improving the accuracy of the driving recorder described in the embodiment of the present invention is applied to a driving recorder, and the driving recorder includes a camera. First, obtain the first light intensity value in the target shooting area corresponding to the camera. When the first light intensity value is less than the light intensity threshold, generate a fill light signal based on the first light intensity value, and then perform fill light on the target shooting area based on the fill light signal to obtain the second light intensity value in the target shooting area. When the second light intensity value is greater than the preset light intensity value, then obtain the first image corresponding to the target shooting area. The preset light intensity value is greater than the light intensity threshold. Determine the first quality evaluation value corresponding to the first image. When the first quality evaluation value is less than the quality evaluation threshold, perform noise reduction processing on the first image to obtain a second image. The second quality evaluation value corresponding to the second image is greater than the quality evaluation threshold. Finally, store the second image in the driving recorder, thereby improving the accuracy of the images recorded by the driving recorder.

[0119] Please refer to Figure 7 , Figure 7 which is a schematic structural diagram of a device for improving the accuracy of a driving recorder provided by an embodiment of the present application. It is applied to a driving recorder, and the driving recorder includes a camera; the device 700 for improving the accuracy of the driving recorder includes: an acquisition unit 701 and a processing unit 702;

[0120] The obtaining unit 701 is configured to obtain a first light intensity value within a target shooting area corresponding to the camera;

[0121] The processing unit 702 is configured to generate a fill light signal based on the first light intensity value when the first light intensity value is less than a light intensity threshold;

[0122] Perform fill light on the target shooting area based on the fill light signal to obtain a second light intensity value within the target shooting area;

[0123] When the second light intensity value is greater than a preset light intensity value, obtain a first image corresponding to the target shooting area; the preset light intensity value is greater than the light intensity threshold;

[0124] Determine a first quality evaluation value corresponding to the first image;

[0125] When the first quality evaluation value is less than a quality evaluation threshold, perform noise reduction processing on the first image to obtain a second image; a second quality evaluation value corresponding to the second image is greater than the quality evaluation threshold;

[0126] Store the second image in the driving recorder.

[0127] In some possible implementation manners, in terms of determining the first quality evaluation value corresponding to the first image, the processing unit 702 is specifically configured to:

[0128] Determine the clarity and noise intensity value corresponding to the first image;

[0129] Determine a first reference quality evaluation value corresponding to the clarity and a second reference quality evaluation value corresponding to the noise intensity value;

[0130] Determine the first quality evaluation value corresponding to the first image based on the first reference quality evaluation value and the second reference quality evaluation value.

[0131] In some possible implementation manners, in terms of determining the clarity and noise intensity value corresponding to the first image, the processing unit 702 is specifically configured to:

[0132] Determine the gradient amplitude corresponding to the first image;

[0133] Obtain a first mapping relationship between the gradient amplitude and the clarity;

[0134] Determine the clarity corresponding to the gradient amplitude based on the first mapping relationship to obtain the clarity corresponding to the first image;

[0135] Divide the first image into n image blocks; n is an integer greater than 1, and the sizes of the n image blocks are the same;

[0136] Determine the reference noise intensity value corresponding to each of the n image blocks to obtain n reference noise intensity values;

[0137] Determine the noise intensity value corresponding to the first image based on the n noise intensity values.

[0138] In some possible implementation manners, in terms of determining the reference noise intensity value corresponding to each of the n image blocks to obtain n reference noise intensity values, the processing unit 702 is specifically configured to:

[0139] Obtain multiple pixel values within a first image block; the first image block is any one of the n image blocks;

[0140] Determine the variance of the pixel values corresponding to the multiple pixel values;

[0141] Obtain a second mapping relationship between the pixel value variance and the reference noise intensity value;

[0142] Determine the reference noise intensity value corresponding to the pixel value variance based on the second mapping relationship to obtain the reference noise intensity value corresponding to the first image block.

[0143] In some possible implementation manners, in terms of determining the first quality assessment value corresponding to the first image based on the first reference quality assessment value and the second reference quality assessment value, the processing unit 702 is specifically configured to:

[0144] Obtain a first reference weight corresponding to the first reference quality assessment value and a second reference weight corresponding to the second reference quality assessment value; the sum of the first reference weight and the second reference weight is 1;

[0145] Obtain the resolution of the first image;

[0146] Determine a first optimization factor corresponding to the resolution;

[0147] Optimize the first reference weight according to the first optimization factor to obtain a first target weight;

[0148] Adjust the second reference weight based on the first target weight to obtain a second target weight; the sum of the first target weight and the second target weight is 1;

[0149] Perform calculations based on the first reference quality assessment value, the second reference quality assessment value, the first target weight, and the second target weight to obtain the first quality assessment value.

[0150] In some possible embodiments, in terms of performing noise reduction processing on the first image to obtain a second image, the processing unit 702 is specifically configured to:

[0151] Obtain historical record images in the driving recorder;

[0152] Perform noise reduction processing on the n image blocks based on the historical record images to obtain n target image blocks;

[0153] Determine the second image based on the n target image blocks;

[0154] Wherein, the performing noise reduction processing on the n image blocks based on the historical record images to obtain n target image blocks includes:

[0155] Determine m similar image blocks corresponding to the first image block in the historical record images;

[0156] Determine the pixel values corresponding to each of the m similar image blocks in the m similar image blocks to obtain m pixel values; m is an integer greater than 1;

[0157] Determine the average pixel value corresponding to the m pixel values;

[0158] Update the pixel value of the first image block to the average pixel value to obtain the target image block corresponding to the first image block.

[0159] In some possible embodiments, in terms of determining m similar image blocks corresponding to the first image block in the historical record images, the processing unit 702 is specifically configured to:

[0160] Divide the historical record images into k candidate image blocks; the k candidate image blocks are all the same size as the first image block; k is an integer greater than m;

[0161] Determine the pixel differences between each of the k candidate image blocks and the first image block to obtain k pixel differences;

[0162] Determine m pixel differences less than a preset pixel difference among the k pixel differences;

[0163] Determine the candidate image blocks corresponding to the m pixel differences to obtain the m similar image blocks.

[0164] Please refer to Figure 8 , Figure 8 is a schematic structural diagram of an electronic device provided by an embodiment of the present application. As Figure 8As shown, the electronic device 800 includes a transceiver 801, a processor 802, and a memory 803. They are connected through a bus 804. The memory 803 is used to store computer programs and data, and the transceiver 801 can transmit the data stored in the memory 803 to the processor 802. The above program includes instructions for performing the following steps:

[0165] Obtain a first light intensity value within a target shooting area corresponding to the camera;

[0166] When the first light intensity value is less than a light intensity threshold, generate a fill light signal based on the first light intensity value;

[0167] Perform fill light on the target shooting area based on the fill light signal to obtain a second light intensity value within the target shooting area;

[0168] When the second light intensity value is greater than a preset light intensity value, obtain a first image corresponding to the target shooting area; the preset light intensity value is greater than the light intensity threshold;

[0169] Determine a first quality evaluation value corresponding to the first image;

[0170] When the first quality evaluation value is less than a quality evaluation threshold, perform noise reduction processing on the first image to obtain a second image; the second quality evaluation value corresponding to the second image is greater than the quality evaluation threshold;

[0171] Store the second image in the driving recorder.

[0172] In some possible implementation manners, in terms of determining the first quality evaluation value corresponding to the first image, the above program includes instructions for performing the following steps:

[0173] Determine the clarity and noise intensity value corresponding to the first image;

[0174] Determine a first reference quality evaluation value corresponding to the clarity and a second reference quality evaluation value corresponding to the noise intensity value;

[0175] Determine the first quality evaluation value corresponding to the first image based on the first reference quality evaluation value and the second reference quality evaluation value.

[0176] In some possible implementation manners, in terms of determining the clarity and noise intensity value corresponding to the first image, the above program includes instructions for performing the following steps:

[0177] Determine the gradient amplitude corresponding to the first image;

[0178] Obtain a first mapping relationship between the gradient amplitude and the clarity;

[0179] Determine the sharpness corresponding to the gradient magnitude based on the first mapping relationship, and obtain the sharpness corresponding to the first image;

[0180] Divide the first image into n image blocks; n is an integer greater than 1, and the sizes of the n image blocks are the same;

[0181] Determine the reference noise intensity value corresponding to each of the n image blocks, and obtain n reference noise intensity values;

[0182] Determine the noise intensity value corresponding to the first image based on the n noise intensity values.

[0183] In some possible implementation manners, in determining the reference noise intensity value corresponding to each of the n image blocks and obtaining n reference noise intensity values, the above program includes instructions for performing the following steps:

[0184] Obtain a plurality of pixel values within a first image block; the first image block is any one of the n image blocks;

[0185] Determine the variance of the pixel values corresponding to the plurality of pixel values;

[0186] Obtain a second mapping relationship between the pixel value variance and the reference noise intensity value;

[0187] Determine the reference noise intensity value corresponding to the pixel value variance based on the second mapping relationship, and obtain the reference noise intensity value corresponding to the first image block.

[0188] In some possible implementation manners, in determining the first quality evaluation value corresponding to the first image based on the first reference quality evaluation value and the second reference quality evaluation value, the above program includes instructions for performing the following steps:

[0189] Obtain a first reference weight corresponding to the first reference quality evaluation value and a second reference weight corresponding to the second reference quality evaluation value; the sum of the first reference weight and the second reference weight is 1;

[0190] Obtain the resolution of the first image;

[0191] Determine a first optimization factor corresponding to the resolution;

[0192] Optimize the first reference weight according to the first optimization factor to obtain a first target weight;

[0193] Adjust the second reference weight based on the first target weight to obtain a second target weight; the sum of the first target weight and the second target weight is 1;

[0194] Calculations are performed based on the first reference quality assessment value, the second reference quality assessment value, the first target weight, and the second target weight to obtain the first quality assessment value.

[0195] In some possible implementation manners, in terms of performing noise reduction processing on the first image to obtain a second image, the above program includes instructions for performing the following steps:

[0196] Obtain the historical record images in the driving recorder;

[0197] Perform noise reduction processing on the n image blocks based on the historical record images to obtain n target image blocks;

[0198] Determine the second image based on the n target image blocks;

[0199] Among them, the performing noise reduction processing on the n image blocks based on the historical record images to obtain n target image blocks includes:

[0200] Determine m similar image blocks corresponding to the first image block in the historical record images;

[0201] Determine the pixel values corresponding to each of the m similar image blocks among the m similar image blocks to obtain m pixel values; m is an integer greater than 1;

[0202] Determine the average pixel value corresponding to the m pixel values;

[0203] Update the pixel value of the first image block to the average pixel value to obtain the target image block corresponding to the first image block.

[0204] In some possible implementation manners, in terms of determining m similar image blocks corresponding to the first image block in the historical record images, the above program includes instructions for performing the following steps:

[0205] Divide the historical record images into k candidate image blocks; the k candidate image blocks are all the same size as the first image block; k is an integer greater than m;

[0206] Determine the pixel differences between each of the k candidate image blocks and the first image block to obtain k pixel differences;

[0207] Determine m pixel differences less than a preset pixel difference among the k pixel differences;

[0208] Determine the candidate image blocks corresponding to the m pixel differences to obtain the m similar image blocks.

[0209] It should be understood that the electronic devices in the present application may include a device for improving the accuracy of a driving recorder, a smart phone (such as an Android phone, an iOS phone, a Windows Phone, etc.), a tablet computer, a handheld computer, a laptop computer, a mobile Internet device MID (Mobile Internet Devices, abbreviated as: MID), or a wearable device, or a server, an edge computing node, etc. The above-mentioned electronic devices are only examples, not an exhaustive list, and include but are not limited to the above-mentioned electronic devices.

[0210] An embodiment of the present application also provides a computer-readable storage medium, which stores a computer program. The computer program is executed by a processor to implement part or all of the steps of any one of the methods for improving the accuracy of a driving recorder described in the above method embodiments.

[0211] An embodiment of the present application also provides a computer program product, which includes a non-transitory computer-readable storage medium storing a computer program. The computer program is operable to cause a computer to execute part or all of the steps of any one of the methods for improving the accuracy of a driving recorder described in the above method embodiments.

[0212] It should be noted that for the foregoing method embodiments, for the sake of simple description, they are all expressed as a series of action combinations. However, those skilled in the art should know that the present application is not limited by the described action sequence, because according to the present application, certain steps can be performed in other sequences or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all optional embodiments, and the actions and modules involved are not necessarily essential to the present application.

[0213] In the above embodiments, the descriptions of the respective embodiments have their own emphases. For the parts not detailed in a certain embodiment, reference may be made to the relevant descriptions of other embodiments.

[0214] In several embodiments provided by the present application, it should be understood that the disclosed device can be implemented in other ways. For example, the device embodiments described above are only illustrative. For example, the division of units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed coupling or direct coupling or communication connection between each other can be through some interfaces. The indirect coupling or communication connection of the device or unit can be in an electrical or other form.

[0215] The unit described as a separate component may or may not be physically separated, and the component shown as a unit may or may not be a physical unit, that is, it may be located in one place or distributed across multiple network units. Some or all of these units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0216] In addition, in each embodiment of this application, each functional unit can be integrated into a processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of a software program module.

[0217] If the integrated unit is implemented in the form of a software program module and sold or used as an independent product, it can be stored in a computer-readable memory. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a memory and includes several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods in each embodiment of this application. And the aforementioned memory includes: USB flash drives, read-only memories (ROM, Read-On ly Memory), random access memories (RAM, Random Access Memory), external hard drives, magnetic disks, or optical discs and other media that can store program codes.

[0218] Those of ordinary skill in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing relevant hardware through a program. This program can be stored in a computer-readable memory, and the memory can include: flash drives, read-only memories (abbreviation: ROM), random access memories (abbreviation: RAM), magnetic disks, or optical discs, etc.

[0219] The above has introduced the embodiments of this application in detail. Specific examples are used in this article to elaborate on the principle and embodiments of this application. The description of the above embodiments is only used to help understand the method and its core idea of this application; at the same time, for those of ordinary skill in the art, according to the idea of this application, there will be changes in the specific embodiments and application scopes. In summary, the content of this specification should not be construed as a limitation to this application.

Claims

1. A method for improving the accuracy of a driving recorder, characterized in that: Applied to a driving recorder, the driving recorder includes a camera; the method includes: Acquire a first light intensity value in a target shooting area corresponding to the camera; When the first light intensity value is less than a light intensity threshold, generating a fill light signal based on the first light intensity value; Fill light is performed on the target shooting area based on the fill light signal to obtain a second light intensity value within the target shooting area; When the second light intensity value is greater than a preset light intensity value, acquiring a first image corresponding to the target shooting area; the preset light intensity value is greater than the light intensity threshold; Determining a first quality assessment value corresponding to the first image; When the first quality assessment value is less than a quality assessment threshold, performing noise reduction processing on the first image to obtain a second image; and a second quality assessment value corresponding to the second image is greater than the quality assessment threshold; The second image is stored in the driving recorder.

2. The method according to claim 1, characterized in that The determining a first quality assessment value corresponding to the first image includes: Determining clarity and noise intensity values ​​corresponding to the first image; Determine a first reference quality assessment value corresponding to the clarity and a second reference quality assessment value corresponding to the noise intensity value; The first quality assessment value corresponding to the first image is determined based on the first reference quality assessment value and the second reference quality assessment value.

3. The method according to claim 2, characterized in that The determining of the clarity and noise intensity values ​​corresponding to the first image includes: determining a gradient magnitude corresponding to the first image; Acquire a first mapping relationship between gradient amplitude and clarity; Determine the definition corresponding to the gradient amplitude based on the first mapping relationship, and obtain the definition corresponding to the first image; Dividing the first image into n image blocks; n is an integer greater than 1, and the n image blocks have the same size; Determine a reference noise intensity value corresponding to each image block in the n image blocks to obtain n reference noise intensity values; A noise intensity value corresponding to the first image is determined based on the n noise intensity values.

4. The method according to claim 3, characterized in that The step of determining a reference noise intensity value corresponding to each of the n image blocks to obtain n reference noise intensity values ​​includes: Acquire multiple pixel values ​​in a first image block; the first image block is any one of the n image blocks; Determining pixel value variances corresponding to the plurality of pixel values; Obtaining a second mapping relationship between pixel value variance and reference noise intensity value; A reference noise intensity value corresponding to the pixel value variance is determined based on the second mapping relationship to obtain a reference noise intensity value corresponding to the first image block.

5. The method according to claim 4, characterized in that The determining the first quality assessment value corresponding to the first image based on the first reference quality assessment value and the second reference quality assessment value includes: Obtaining a first reference weight corresponding to the first reference quality assessment value and a second reference weight corresponding to the second reference quality assessment value; the sum of the first reference weight and the second reference weight is 1; Acquire the resolution of the first image; determining a first optimization factor corresponding to the resolution; Optimize the first reference weight according to the first optimization factor to obtain a first target weight; Adjust the second reference weight based on the first target weight to obtain a second target weight; the sum of the first target weight and the second target weight is 1; The first quality evaluation value is obtained by performing calculation based on the first reference quality evaluation value, the second reference quality evaluation value, the first target weight, and the second target weight.

6. The method according to claim 5, characterized in that The performing noise reduction processing on the first image to obtain a second image includes: Acquire the historical record image in the driving recorder; Performing noise reduction processing on the n image blocks based on the historical record images to obtain n target image blocks; determining the second image based on the n target image blocks; The step of performing noise reduction processing on the n image blocks based on the historical record image to obtain n target image blocks includes: Determine m similar image blocks in the historical record image corresponding to the first image block; Determine the pixel value corresponding to each similar image block in the m similar image blocks to obtain m pixel values, where m is an integer greater than 1; Determine an average pixel value corresponding to the m pixel values; The pixel value of the first image block is updated to the average pixel value to obtain a target image block corresponding to the first image block.

7. The method according to claim 6, characterized in that The determining m similar image blocks in the historical record image corresponding to the first image block includes: Divide the historical record image into k candidate image blocks; the k candidate image blocks are all the same size as the first image block; k is an integer greater than m; Determine a pixel difference between each image block in the k candidate image blocks and the first image block to obtain k pixel difference values; Determine m pixel differences among the k pixel differences that are smaller than a preset pixel difference; Determine the candidate image blocks corresponding to the m pixel differences to obtain the m similar image blocks.

8. A device for improving the accuracy of a driving recorder, characterized in that: Applied to a driving recorder, the driving recorder includes a camera; the driving recorder accuracy improvement device includes: an acquisition unit and a processing unit; The acquisition unit is used to acquire a first light intensity value within a target shooting area corresponding to the camera; The processing unit is configured to generate a fill light signal based on the first light intensity value when the first light intensity value is less than a light intensity threshold; Fill light is performed on the target shooting area based on the fill light signal to obtain a second light intensity value within the target shooting area; When the second light intensity value is greater than a preset light intensity value, acquiring a first image corresponding to the target shooting area; the preset light intensity value is greater than the light intensity threshold; Determining a first quality assessment value corresponding to the first image; When the first quality assessment value is less than a quality assessment threshold, performing noise reduction processing on the first image to obtain a second image; and a second quality assessment value corresponding to the second image is greater than the quality assessment threshold; The second image is stored in the driving recorder.

9. An electronic device, characterized in that: The method comprises a processor, a memory, a communication interface and one or more programs, wherein the one or more programs are stored in the memory and configured to be executed by the processor, and the one or more programs include instructions for executing the steps in the method described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and the computer program is executed by a processor to implement the method according to any one of claims 1 to 7.

Images