Vehicle-mounted driving auxiliary image quality optimization method and device

Abstract

The invention discloses a vehicle-mounted driving auxiliary image quality optimization method and device. The method comprises the following steps: acquiring image stream data of a vehicle-mounted driving auxiliary image; performing image enhancement and illumination compensation processing on an image in the image stream data based on a homomorphic filtering algorithm to obtain an image after homomorphic filtering; carrying out image gray scale adjustment on the image after homomorphic filtering based on a gamma transformation algorithm to obtain an image after gamma transformation; and carrying out denoising and detail enhancement processing on the image after gamma transformation based on a guiding filtering algorithm, and outputting an optimized vehicle-mounted driving auxiliary image. According to the method, the noise reduction effect can be achieved, meanwhile, the edge and other detail parts of the image can be well reserved, the utilization rate of the GPU is considered, and the image quality optimization efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the field of autonomous driving technology, and in particular to a method and device for optimizing the image quality of a vehicle-mounted driving assistance image. Background Art

[0002] With the continuous advancement of the automotive industry and intelligent driving technology, on-board driving assistance imaging systems have become standard in modern cars. These systems use multiple cameras to collect images of the vehicle's surroundings in real time to assist the driver in making safe driving decisions. However, complex and changing environmental factors during driving often lead to a decline in image quality, affecting the effectiveness and reliability of driving assistance images. Among them, the main influencing factors include: lighting changes, weather conditions, motion blur, and camera performance limitations.

[0003] Currently, the image quality optimization technologies for vehicle-mounted driving assistance images mainly include the following:

[0004] 1) Image denoising: Use Gaussian filtering, median filtering and other methods to reduce the image quality degradation caused by sensor noise and environmental noise.

[0005] 2) Image enhancement: Through histogram equalization, adaptive histogram equalization and other technologies, the overall brightness and contrast of the image are improved to highlight important details.

[0006] 3) Super-resolution reconstruction: Use deep learning techniques (such as convolutional neural networks) to reconstruct low-resolution images to obtain high-resolution and high-detail images.

[0007] 4) Bad weather processing: Through defogging, deraining and other algorithms, the interference caused by bad weather in the image is removed to improve the clarity of the image.

[0008] 5) Color correction: adjust white balance and color balance, correct image color distortion, and ensure color authenticity and consistency.

[0009] Although these technologies have improved the quality of driving assistance images to a certain extent, they still face some challenges. For example, the stability and robustness of the image quality optimization algorithm under different lighting and weather conditions still need to be further improved. In addition, in order to meet the real-time requirements of the driving environment, the computational efficiency of the algorithm also needs to be optimized.

[0010] Therefore, the prior art needs to be improved. Summary of the invention

[0011] The technical problem to be solved by the present invention is that, in view of the defects of the prior art, the present invention provides a method and device for optimizing the image quality of vehicle-mounted driving auxiliary images, so as to solve the problems of low stability and low computational efficiency of the traditional image quality optimization algorithm of vehicle-mounted driving auxiliary images.

[0012] The technical solution adopted by the present invention to solve the technical problem is as follows:

[0013] In a first aspect, the present invention provides a method for optimizing the image quality of a vehicle-mounted driving assistance image, comprising:

[0014] Obtain image stream data of vehicle-mounted driving assistance images;

[0015] Performing image enhancement and illumination compensation processing on the image in the image stream data based on a homomorphic filtering algorithm to obtain an image after homomorphic filtering;

[0016] Performing image grayscale adjustment on the homomorphically filtered image based on a gamma transformation algorithm to obtain a gamma transformed image;

[0017] Based on the guided filtering algorithm, the gamma-transformed image is subjected to denoising and detail enhancement processing, and an optimized vehicle-mounted driving auxiliary image is output.

[0018] In one implementation, the step of acquiring image stream data of a vehicle-mounted driving assistance image includes:

[0019] Acquire multiple frames of continuous driving assistance images collected by a vehicle-mounted camera to obtain the image stream data; wherein the images in the image stream data include: any one of RGB images, YUV images and YCrCb images.

[0020] In one implementation, performing image enhancement and illumination compensation processing on the image in the image stream data based on a homomorphic filtering algorithm to obtain a homomorphic filtered image includes:

[0021] Performing logarithmic transformation processing on the images in the image stream data;

[0022] Suppressing low-frequency components of the image processed by logarithmic transformation through a high-pass filter and enhancing high-frequency components of the image processed by logarithmic transformation to obtain a filtered result;

[0023] Performing exponential transformation on the filtered result to obtain the homomorphically filtered image.

[0024] In one implementation, the formula of the homomorphic filtering algorithm is:

[0025]

[0026] Among them, γH and γ L They represent the weighted components of high frequency and low frequency of the image, respectively, and c represents the threshold for identifying whether it is high frequency;

[0027] Represents Gaussian blur or Butterworth blur on the image.

[0028] In one implementation, the grayscale adjustment of the homomorphically filtered image based on a gamma transformation algorithm to obtain a gamma transformed image includes:

[0029] Convert the pixel values ​​of the homomorphically filtered image into real numbers between 0 and 1;

[0030] According to the formula of gamma transformation, the corresponding value of the pixel normalized data with 1 / γ as the exponent is calculated, and the normalized data is pre-compensated;

[0031] The pre-compensated real value is inversely transformed into a value between 0 and 255 to obtain the gamma-transformed image.

[0032] In one implementation, the denoising and detail enhancement processing is performed on the gamma-transformed image based on a guided filtering algorithm to output an optimized vehicle-mounted driving assistance image, including:

[0033] Determine the sliding window size and denoising strength hyperparameters for mean filtering;

[0034] The gamma-transformed image is subjected to denoising and detail enhancement processing according to the sliding window size of the mean filter and the denoising intensity hyperparameter, and an optimized vehicle-mounted driving auxiliary image is output.

[0035] In one implementation, determining the sliding window size and denoising strength hyperparameters of the mean filter includes:

[0036] According to the GPU utilization and the denoising effect, the sliding window size of the mean filter and the denoising strength hyperparameters are determined respectively.

[0037] In a second aspect, the present invention provides a vehicle-mounted driving auxiliary image quality optimization device, comprising:

[0038] An image acquisition module is used to acquire image stream data of vehicle-mounted driving assistance images;

[0039] A homomorphic filtering module, used for performing image enhancement and illumination compensation processing on the image in the image stream data based on a homomorphic filtering algorithm to obtain an image after homomorphic filtering;

[0040] An image transformation module, used for adjusting the grayscale of the homomorphically filtered image based on a gamma transformation algorithm to obtain a gamma transformed image;

[0041] The guided filtering module is used to perform denoising and detail enhancement on the gamma-transformed image based on a guided filtering algorithm, and output an optimized vehicle-mounted driving auxiliary image.

[0042] In a third aspect, the present invention provides a terminal, comprising: a processor and a memory, wherein the memory stores a vehicle-mounted driving auxiliary image image quality optimization program, and when the vehicle-mounted driving auxiliary image image quality optimization program is executed by the processor, it is used to implement the operation of the vehicle-mounted driving auxiliary image image quality optimization method as described in the first aspect.

[0043] In a fourth aspect, the present invention further provides a medium, which is a computer-readable storage medium, and which stores a vehicle-mounted driving assistance image quality optimization program. When the vehicle-mounted driving assistance image quality optimization program is executed by a processor, it is used to implement the operation of the vehicle-mounted driving assistance image quality optimization method as described in the first aspect.

[0044] The present invention adopts the above technical solution to achieve the following effects:

[0045] The present invention can perform image enhancement and illumination compensation processing on the image in the image stream data based on the homomorphic filtering algorithm by acquiring the image stream data of the vehicle-mounted driving auxiliary image, and adjust the image grayscale of the homomorphic filtered image based on the gamma transformation algorithm, so that the gamma transformed image can be denoised and enhanced in detail based on the guided filtering algorithm, and the optimized vehicle-mounted driving auxiliary image can be output. The present invention can achieve the effect of noise reduction, and at the same time can well retain the edges and other details of the image, taking into account the utilization rate of the GPU, and improving the efficiency of image quality optimization. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the structures shown in these drawings without paying creative work.

[0047] Figure 1 It is a flow chart of a method for optimizing the image quality of vehicle-mounted driving assistance images in one implementation of the present invention.

[0048] Figure 2 It is a comparative schematic diagram of the vehicle-mounted driving assistance image before and after optimization in one implementation of the present invention.

[0049] Figure 3 It is a functional principle diagram of a terminal in one implementation of the present invention.

[0050] The purpose, features and advantages of the present invention will be further described with reference to the accompanying drawings in conjunction with the embodiments. DETAILED DESCRIPTION

[0051] In order to make the purpose, technical solution and advantages of the present invention clearer and more specific, the present invention is further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0052] Exemplary Methods

[0053] At present, the image quality optimization technologies for vehicle-mounted driving assistance images mainly include the following: image denoising, image enhancement, super-resolution reconstruction, severe weather processing, color correction and other image quality optimization methods; although these technologies have improved the quality of driving assistance images to a certain extent, they still face some challenges. For example, under different lighting and weather conditions, the stability and robustness of the image quality optimization algorithm still need to be further improved. In addition, in order to meet the real-time requirements of the driving environment, the computational efficiency of the algorithm also needs to be optimized.

[0054] In view of the above technical problems, an embodiment of the present invention provides a method for optimizing the image quality of an on-board driving auxiliary image. The method acquires the image stream data of the on-board driving auxiliary image, and can perform image enhancement and illumination compensation processing on the image in the image stream data based on the homomorphic filtering algorithm, and adjust the image grayscale of the homomorphic filtered image based on the gamma transformation algorithm, so that the gamma transformed image can be denoised and enhanced in detail based on the guided filtering algorithm, and the optimized on-board driving auxiliary image can be output. The embodiment of the present invention can achieve the effect of noise reduction, while well preserving the edges and other details of the image, taking into account the utilization rate of the GPU and improving the efficiency of image quality optimization.

[0055] like Figure 1 As shown, an embodiment of the present invention provides a method for optimizing the image quality of a vehicle-mounted driving assistance image, comprising the following steps:

[0056] Step S100, acquiring image stream data of the vehicle-mounted driving assistance image.

[0057] In this embodiment, after investigating the existing mass-produced models on the market and the experience in the development process of the corresponding functions, it is found that the products that have been mass-produced on the existing market usually do not use high-cost high-definition cameras, but use lower-cost automotive-grade cameras. The pictures output by these cameras usually have a lot of ISP (Image Signal Processor) noise points. At the same time, the darker the environment, the more obvious these noise points are.

[0058] Through the analysis of ISP noise points, the denoising process of this kind of noise can be approximately equivalent to a dehazing operation. At the same time, the details of the image need to be retained, so the high-frequency part needs to be selectively enhanced or retained. Finally, it is necessary to take into account the adaptability of the image quality optimization algorithm and the utilization rate of the GPU (that is, the computing power cost), and improve the efficiency of image quality optimization while ensuring the image quality optimization effect.

[0059] In this embodiment, a combined pipeline of adaptive image optimization operators is proposed, which can achieve the effect of noise reduction while well preserving the edges and other details of the image; moreover, it can also well take into account the utilization rate of the GPU. The combined pipeline in this embodiment refers to a GPU rendering pipeline for image optimization based on operators of homomorphic filtering, gamma transformation and guided filtering.

[0060] In this embodiment, before optimizing the image quality of the vehicle-mounted driving assistance image, it is necessary to obtain the image stream data of the vehicle-mounted driving assistance image to perform homomorphic filtering, gamma transformation and guided filtering on the image stream data of the vehicle-mounted driving assistance image.

[0061] Specifically, in an implementation of this embodiment, step S100 includes the following steps:

[0062] Step S101, acquiring multiple frames of continuous driving assistance images captured by a vehicle-mounted camera to obtain the image stream data.

[0063] In this embodiment, the vehicle-mounted driving auxiliary image refers to the image of the vehicle's surrounding environment displayed by the vehicle-mounted camera, which provides convenience and improves safety for parking and reversing. The vehicle-mounted driving auxiliary image is divided into reversing image, 360-degree panoramic image, vehicle side blind spot image, 180-degree transparent chassis image and 360-degree transparent chassis image; these vehicle-mounted driving auxiliary images can be collected by one or more vehicle-mounted cameras set in the rear trunk cover, bumper, front, rear, under the rearview mirrors on both sides and the vehicle chassis.

[0064] In this embodiment, image stream data can be obtained by acquiring multiple frames of continuous driving assistance images captured by the above-mentioned vehicle-mounted cameras; in these image stream data, the image includes: any one of RGB image, YUV image and YCrCb image; that is, the image format in the image stream data can be RGB format, YUV format, or YCrCb format.

[0065] like Figure 1 As shown, in one implementation of the embodiment of the present invention, the method for optimizing the image quality of the vehicle-mounted driving assistance image further includes the following steps:

[0066] Step S200: performing image enhancement and illumination compensation processing on the image in the image stream data based on a homomorphic filtering algorithm to obtain an image after homomorphic filtering.

[0067] In this embodiment, the acquired image stream data is used as input, and the image in the image stream data is subjected to image enhancement and illumination compensation processing by a homomorphic filtering algorithm to obtain a homomorphic filtered image; the homomorphic filtering algorithm is an algorithm for image enhancement and illumination compensation, and the algorithm is intended to simultaneously process the reflection component and the illumination component of the image. Homomorphic filtering can be achieved by converting the image into a logarithmic domain, separating and compressing the illumination component, and enhancing the reflection component.

[0068] Specifically, in an implementation of this embodiment, step S200 includes the following steps:

[0069] Step S201, performing logarithmic transformation processing on the image in the image stream data;

[0070] Step S202, suppressing the low-frequency components of the image processed by logarithmic transformation through a high-pass filter, and enhancing the high-frequency components of the image processed by logarithmic transformation, to obtain a filtered result;

[0071] Step S203, performing exponential transformation on the filtered result to obtain the homomorphically filtered image.

[0072] In this embodiment, the specific steps of the homomorphic filtering algorithm include:

[0073] First, logarithmic transformation is performed on the images in the image stream data to convert the images into a logarithmic domain.

[0074] Then, a high-pass filter is used to suppress low-frequency components and enhance high-frequency components to achieve homomorphic filtering of the image. In the process of high-pass filter suppression, the suppressed low-frequency components are the illumination components of the image. Because this part contains overall brightness changes or background noise, it needs to be suppressed by appropriate methods to improve image quality or signal readability and processability. Therefore, the high-pass filter can reduce the impact of illumination by filtering out low-frequency components in the image. The enhanced high-frequency components are the reflection components of the image. This part usually includes the reflective part of the image, which is very important for improving the detail clarity and perceived quality of the image. Therefore, the high-pass filter can filter out the low-frequency components in the image, thereby highlighting the high-frequency details.

[0075] Finally, the filtered result is subjected to exponential transformation to restore the image, thereby obtaining the homomorphically filtered image.

[0076] In this embodiment, the formula of the homomorphic filtering algorithm is:

[0077]

[0078] Among them, γ H and γ L They represent the weighted components of high frequency and low frequency of the image, respectively, and c represents the threshold for identifying whether it is high frequency;

[0079] Represents Gaussian blur or Butterworth blur on the image.

[0080] As an example, in this embodiment, when the homomorphic filtering algorithm is used for filtering, γ H = 0.9 and γ L =0.15, c=120 is in line with the expected homomorphic filtering effect; at the same time, in the process of Gaussian blur or Butterworth blur on the image, Gaussian blur is preferably used as a high-pass filter to suppress low-frequency (light) components and enhance high-frequency (reflection) components.

[0081] It can be understood that in the process of suppressing low-frequency (light) components, methods such as histogram equalization and local contrast enhancement can also be used to suppress low-frequency (light) components; and in the process of enhancing high-frequency (reflection) components, methods such as edge enhancement, detail enhancement and reflection enhancement can also be used to enhance the high-frequency (reflection) components.

[0082] In this embodiment, the image is filtered by homomorphic filtering, which can effectively reduce the image quality problem caused by uneven ambient lighting and make the details of the image clearer and more uniform.

[0083] like Figure 1As shown, in one implementation of the embodiment of the present invention, the method for optimizing the image quality of the vehicle-mounted driving assistance image further includes the following steps:

[0084] Step S300, adjusting the grayscale of the homomorphically filtered image based on a gamma transformation algorithm to obtain a gamma transformed image.

[0085] In this embodiment, after image enhancement and illumination compensation processing are performed on the image in the image stream data, the image grayscale is adjusted on the homomorphic filtered image based on a gamma transformation algorithm to obtain a gamma transformed image; wherein the gamma transformation is a nonlinear operation that adjusts the grayscale of the image to make the image more visually natural.

[0086] Specifically, in an implementation of this embodiment, step S300 includes the following steps:

[0087] Step S301, converting the pixel values ​​of the image after homomorphic filtering into real numbers between 0 and 1;

[0088] Step S302, according to the formula of gamma transformation, the corresponding value of the pixel normalized data with 1 / γ as the exponent is calculated, and the normalized data is pre-compensated;

[0089] Step S303, inversely transforming the pre-compensated real value into a value between 0 and 255 to obtain the image after the gamma transformation.

[0090] In this embodiment, the steps of gamma transformation are:

[0091] First, the image after homomorphic filtering is normalized: the pixel values ​​are converted into real numbers between 0 and 1. The calculation method is as follows:

[0092] (value+0.5) / 256;

[0093] Secondly, the normalized data is pre-compensated: According to the formula of gamma transformation, the corresponding value of the pixel normalized data with 1 / γ as the exponent is calculated, and pre-compensation is performed based on the calculated corresponding value.

[0094] Finally, the pre-compensated real value is denormalized: the pre-compensated real value is denormalized to a value between 0 and 255.

[0095] As an example, in this embodiment, 3.0 to 4.0 is selected as the value of value.

[0096] In this embodiment, the image grayscale is adjusted on the homomorphically filtered image by using a gamma transformation algorithm, so that the contrast and color vividness of the image can be enhanced, thereby improving the display effect of the image captured under low light conditions.

[0097] like Figure 1 As shown, in one implementation of the embodiment of the present invention, the method for optimizing the image quality of the vehicle-mounted driving assistance image further includes the following steps:

[0098] Step S400: De-noising and detail enhancement processing is performed on the gamma-transformed image based on a guided filtering algorithm, and an optimized vehicle-mounted driving auxiliary image is output.

[0099] In this embodiment, after the homomorphic filtered image is gamma transformed, the guided filtering algorithm can be used to perform denoising and detail enhancement on the gamma transformed image to achieve detail enhancement in the image; a guided filter is a filter based on a local linear model, which combines a guide image and an input image, and can effectively perform image enhancement, denoising and edge preservation.

[0100] Specifically, in an implementation of this embodiment, step S400 includes the following steps:

[0101] Step S401, determining the sliding window size of the mean filter and the denoising strength hyperparameters;

[0102] In one implementation of this embodiment, step S401 includes the following steps:

[0103] Step S401a, determining the sliding window size of the mean filter and the denoising strength hyperparameters according to the GPU utilization and the denoising effect.

[0104] Step S402, denoising and detail enhancement processing is performed on the gamma-transformed image according to the sliding window size of the mean filter and the denoising intensity hyperparameter, and an optimized vehicle-mounted driving auxiliary image is output.

[0105] In this embodiment, the core idea of ​​guided filtering is to use a guiding image to guide the filter to smooth the input image. The guiding image is usually a high-resolution version of the input image or other information related to the input image. The guided filter calculates each pixel value in the output image by the following formula:

[0106] I'(x)=a(x)·I(x)+b(x);

[0107] Where I'(x) represents the pixel value of the output image, I(x) represents the pixel value of the input image, a(x) and b(x) represent the weights and bias terms calculated based on the guide image, respectively. These weights and bias terms are calculated by convolving the input image and the guide image within the window using the guided filter.

[0108] In this embodiment, first, the image after gamma transformation is used as the input image of the guided filtering, and then a high-resolution version of the input image or some image related to the input image is selected as the guiding image.

[0109] After selecting the guide image, it is also necessary to determine the sliding window size of the mean filter and the denoising strength hyperparameter; wherein the sliding window size (window size) is the sliding window size of the mean filter, and the sliding window size can be selected by comprehensively considering the GPU utilization and the denoising effect during the selection process, for example, window_size=3 is selected as the sliding window size; and the denoising strength hyperparameter (eps) can be selected by comprehensively considering the GPU utilization and the denoising effect during the selection process, for example, eps=0.001~0.002 is used as the hyperparameter size.

[0110] After selecting the sliding window size and denoising strength hyperparameters, the mean of the input image and the guide image of the guided filter in the sliding window is calculated, and the covariance of the input image and the guide image of the guided filter after normalization is calculated; then, based on the mean and covariance calculated above, the coefficients a and b of the mean filter are calculated, and these coefficients are used to smooth and enhance the pixel values ​​of the output image. Finally, the input image is filtered using the calculated coefficients a and b of the mean filter to obtain the output image.

[0111] This embodiment performs denoising and detail enhancement processing on the gamma-transformed image through a guided filtering algorithm, which can retain the structural information and details of the image while effectively removing noise; after the gamma-transformed image is guided filtered, an optimized vehicle-mounted driving auxiliary image can be output.

[0112] It is worth mentioning that in other implementations of this embodiment, the image quality optimization of the vehicle-mounted driving assistance image can also be achieved by the following methods:

[0113] For example, the image quality of the vehicle-mounted driving assistance image can be optimized by adjusting the underlying ISP (Image Signal Processing), but this method will increase additional costs, especially if there are modules using deep learning algorithms in the vehicle-mounted assisted driving system. Changing the ISP of the image means that the data of the new ISP image needs to be re-collected; or the image quality can be optimized using deep learning methods, but this method brings high computing power costs and may not meet the real-time requirements of vehicle-mounted driving images.

[0114] In this embodiment, through the combination of homomorphic filtering + gamma transformation + guided filtering operators, the problem of uneven illumination can be firstly handled by homomorphic filtering, then the overall brightness and contrast can be adjusted by gamma transformation, and finally the guided filtering can be used for fine processing, thereby ensuring the provision of high-quality driving auxiliary images in various complex environments. At the same time, these three operators have corresponding efficient implementation methods, which can be deployed efficiently and quickly on the vehicle side, and the algorithm overhead is relatively small.

[0115] like Figure 2 As shown, Figure 2 This is a comparison diagram of the on-board driving assistance image before and after optimization. The final display results, the images on the left (upper left and lower left) are the images before the algorithm is applied, and the images on the right (upper right and lower right) are the images after the algorithm is applied. It can be seen that the algorithm has a good denoising effect and retains complete details, and has a certain compensation for dark scene lighting.

[0116] This embodiment achieves the following technical effects through the above technical solution:

[0117] By acquiring the image stream data of the vehicle-mounted driving auxiliary image, this embodiment can perform image enhancement and illumination compensation processing on the image in the image stream data based on the homomorphic filtering algorithm, and adjust the image grayscale of the homomorphic filtered image based on the gamma transformation algorithm, so that the gamma transformed image can be denoised and enhanced in detail based on the guided filtering algorithm, and the optimized vehicle-mounted driving auxiliary image can be output. The solution adopted in this embodiment can achieve the effect of noise reduction, while well preserving the edges and other details of the image, taking into account the utilization rate of the GPU and improving the efficiency of image quality optimization.

[0118] Exemplary Devices

[0119] Based on the above embodiments, the present invention further provides a vehicle-mounted driving auxiliary image quality optimization device, comprising:

[0120] An image acquisition module is used to acquire image stream data of vehicle-mounted driving assistance images;

[0121] A homomorphic filtering module, used for performing image enhancement and illumination compensation processing on the image in the image stream data based on a homomorphic filtering algorithm to obtain an image after homomorphic filtering;

[0122] An image transformation module, used for adjusting the grayscale of the homomorphically filtered image based on a gamma transformation algorithm to obtain a gamma transformed image;

[0123] The guided filtering module is used to perform denoising and detail enhancement on the gamma-transformed image based on a guided filtering algorithm, and output an optimized vehicle-mounted driving auxiliary image.

[0124] This embodiment achieves the following technical effects through the above technical solution:

[0125] By acquiring the image stream data of the vehicle-mounted driving auxiliary image, this embodiment can perform image enhancement and illumination compensation processing on the image in the image stream data based on the homomorphic filtering algorithm, and adjust the image grayscale of the homomorphic filtered image based on the gamma transformation algorithm, so that the gamma transformed image can be denoised and enhanced in detail based on the guided filtering algorithm, and the optimized vehicle-mounted driving auxiliary image can be output. The solution adopted in this embodiment can achieve the effect of noise reduction, while well preserving the edges and other details of the image, taking into account the utilization rate of the GPU and improving the efficiency of image quality optimization.

[0126] Based on the above embodiment, the present invention further provides a terminal, whose principle block diagram can be as follows: Figure 3 shown.

[0127] The terminal includes: a processor, a memory, an interface, a display screen and a communication module connected through a system bus; wherein the processor of the terminal is used to provide computing and control capabilities; the memory of the terminal includes a storage medium and an internal memory; the storage medium stores an operating system and a computer program; the internal memory provides an environment for the operation of the operating system and the computer program in the storage medium; the interface is used to connect to external devices; the display screen is used to display corresponding information; and the communication module is used to communicate with a cloud server or other devices.

[0128] When the computer program is executed by a processor, it is used to implement the operation of the method for optimizing the image quality of vehicle-mounted driving assistance images.

[0129] It can be understood by those skilled in the art that Figure 3 The principle block diagram shown in the figure is only a block diagram of a partial structure related to the scheme of the present invention, and does not constitute a limitation on the terminal to which the scheme of the present invention is applied. The specific terminal may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0130] In one embodiment, a terminal is provided, which includes: a processor and a memory, the memory storing a vehicle-mounted driving assistance image quality optimization program, and the vehicle-mounted driving assistance image quality optimization program is used to implement the operation of the above-mentioned vehicle-mounted driving assistance image quality optimization method when executed by the processor.

[0131] In one embodiment, a storage medium is provided, wherein the storage medium stores a vehicle-mounted driving assistance image quality optimization program, and when the vehicle-mounted driving assistance image quality optimization program is executed by a processor, it is used to implement the operation of the above-mentioned vehicle-mounted driving assistance image quality optimization method.

[0132] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing related hardware through a computer program, and the computer program can be stored in a non-volatile storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided by the present invention can include non-volatile and volatile memory.

[0133] In summary, the present invention provides a method and device for optimizing the image quality of vehicle-mounted driving auxiliary images, the method comprising: obtaining image stream data of vehicle-mounted driving auxiliary images; performing image enhancement and illumination compensation processing on the images in the image stream data based on a homomorphic filtering algorithm to obtain a homomorphic filtered image; performing image grayscale adjustment on the homomorphic filtered image based on a gamma transformation algorithm to obtain a gamma transformed image; performing denoising and detail enhancement processing on the gamma transformed image based on a guided filtering algorithm to output an optimized vehicle-mounted driving auxiliary image. The present invention can achieve the effect of noise reduction, while well preserving the edges and other details of the image, taking into account the utilization rate of the GPU and improving the efficiency of image quality optimization.

[0134] It should be understood that the application of the present invention is not limited to the above examples. For ordinary technicians in this field, improvements or changes can be made based on the above description. All these improvements and changes should fall within the scope of protection of the claims attached to the present invention.

Claims

1. A method for optimizing the image quality of vehicle-mounted driving assistance images, characterized in that: include: Obtain image stream data of vehicle-mounted driving assistance images; Performing image enhancement and illumination compensation processing on the image in the image stream data based on a homomorphic filtering algorithm to obtain an image after homomorphic filtering; Performing image grayscale adjustment on the homomorphically filtered image based on a gamma transformation algorithm to obtain a gamma transformed image; Based on the guided filtering algorithm, the gamma-transformed image is subjected to denoising and detail enhancement processing, and an optimized vehicle-mounted driving auxiliary image is output.

2. The method for optimizing the image quality of vehicle-mounted driving assistance images according to claim 1, characterized in that: The step of acquiring image stream data of the vehicle-mounted driving assistance image comprises: Acquire multiple frames of continuous driving assistance images collected by a vehicle-mounted camera to obtain the image stream data; wherein the images in the image stream data include: any one of RGB images, YUV images and YCrCb images.

3. The method for optimizing the image quality of vehicle-mounted driving assistance images according to claim 1, characterized in that: The performing image enhancement and illumination compensation processing on the image in the image stream data based on the homomorphic filtering algorithm to obtain the homomorphic filtered image includes: Performing logarithmic transformation processing on the images in the image stream data; Suppressing low-frequency components of the image processed by logarithmic transformation through a high-pass filter and enhancing high-frequency components of the image processed by logarithmic transformation to obtain a filtered result; Performing exponential transformation on the filtered result to obtain the homomorphically filtered image.

4. The method for optimizing the image quality of vehicle-mounted driving assistance images according to claim 3, characterized in that: The formula of the homomorphic filtering algorithm is: Among them, γ H and γ L They represent the weighted components of high frequency and low frequency of the image, respectively, and c represents the threshold for identifying whether it is high frequency; Represents Gaussian blur or Butterworth blur on the image.

5. The method for optimizing the image quality of vehicle-mounted driving assistance images according to claim 1, characterized in that: The step of adjusting the grayscale of the homomorphically filtered image based on a gamma transformation algorithm to obtain a gamma transformed image includes: Convert the pixel values ​​of the homomorphically filtered image into real numbers between 0 and 1; According to the formula of gamma transformation, the corresponding value of the pixel normalized data with 1 / γ as the exponent is calculated, and the normalized data is pre-compensated; The pre-compensated real value is inversely transformed into a value between 0 and 255 to obtain the gamma-transformed image.

6. The method for optimizing the image quality of vehicle-mounted driving assistance images according to claim 1, characterized in that: The method of performing denoising and detail enhancement processing on the gamma-transformed image based on the guided filtering algorithm and outputting an optimized vehicle-mounted driving auxiliary image comprises: Determine the sliding window size and denoising strength hyperparameters for mean filtering; The gamma-transformed image is subjected to denoising and detail enhancement processing according to the sliding window size of the mean filter and the denoising intensity hyperparameter, and an optimized vehicle-mounted driving auxiliary image is output.

7. The method for optimizing the image quality of vehicle-mounted driving assistance images according to claim 6, characterized in that: The step of determining the sliding window size and denoising strength hyperparameters of the mean filter includes: According to the GPU utilization and the denoising effect, the sliding window size of the mean filter and the denoising strength hyperparameters are determined respectively.

8. A vehicle-mounted driving auxiliary image quality optimization device, characterized in that: include: An image acquisition module is used to acquire image stream data of vehicle-mounted driving assistance images; A homomorphic filtering module, used for performing image enhancement and illumination compensation processing on the image in the image stream data based on a homomorphic filtering algorithm to obtain an image after homomorphic filtering; An image transformation module, used for adjusting the grayscale of the homomorphically filtered image based on a gamma transformation algorithm to obtain a gamma transformed image; The guided filtering module is used to perform denoising and detail enhancement on the gamma-transformed image based on a guided filtering algorithm, and output an optimized vehicle-mounted driving auxiliary image.

9. A terminal, characterized in that: include: A processor and a memory, wherein the memory stores a vehicle-mounted driving assistance image quality optimization program, and when the vehicle-mounted driving assistance image quality optimization program is executed by the processor, it is used to implement the operation of the vehicle-mounted driving assistance image quality optimization method as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a vehicle-mounted driving assistance image quality optimization program, and when the vehicle-mounted driving assistance image quality optimization program is executed by the processor, it is used to implement the operation of the vehicle-mounted driving assistance image quality optimization method as described in any one of claims 1-7.

Images