An image processing method, apparatus, device and medium

Abstract

This invention discloses an image processing method, apparatus, device, and medium. The method includes: acquiring vehicle usage scenario parameters corresponding to a surrounding environment image to be processed; determining a target feature acquisition model corresponding to the vehicle usage scenario of the surrounding environment image from a pre-trained pool of candidate feature acquisition models based on the vehicle usage scenario parameters; inputting the surrounding environment image to be processed into the target feature acquisition model to obtain bird's-eye view features corresponding to the surrounding environment image; the input of the target feature acquisition model is the surrounding environment image, and the output is the bird's-eye view features corresponding to the surrounding environment image. This invention allows the determination of a feature acquisition model corresponding to the vehicle usage scenario of the surrounding environment image based on the vehicle usage scenario parameters, enabling the rapid and accurate acquisition of bird's-eye view features based on the corresponding feature acquisition model.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of computer, and in particular, to an image processing method, device, equipment and medium. BACKGROUND

[0002] In the use process of the vehicle, the left front view camera arranged on the vehicle collects images in the left front view direction of the vehicle, the front view camera arranged on the vehicle collects images in the front view direction of the vehicle, the right front view camera arranged on the vehicle collects images in the right front view direction of the vehicle, the left rear view camera arranged on the vehicle collects images in the left rear view direction of the vehicle, the rear view camera arranged on the vehicle collects images in the rear view direction of the vehicle, and the right rear view camera arranged on the vehicle collects images in the right rear view direction of the vehicle. The images in the left front view direction, the front view direction, the right front view direction, the left rear view direction, the rear view direction and the right rear view direction of the vehicle constitute the surrounding environment image of the vehicle. After obtaining the surrounding environment image of the vehicle, the surrounding environment image is processed through a preset feature acquisition model to obtain the bird's eye view feature corresponding to the surrounding environment image.

[0003] In the related art, a commonly used image processing scheme for the surrounding environment image of the vehicle is as follows: the surrounding environment image to be processed is input into a fixedly set feature acquisition model, and the bird's eye view feature corresponding to the surrounding environment image output by the feature acquisition model is obtained. In the use process of the vehicle, the vehicle is in different vehicle use scenarios. The surrounding environment images in different vehicle use scenarios are obviously different, and the bird's eye view features corresponding to the surrounding environment images in different vehicle use scenarios are also obviously different. In the image processing scheme in the related art, the fixedly set feature acquisition model cannot process the surrounding environment images in different vehicle use scenarios, and cannot guarantee the accuracy of the bird's eye view feature extraction. SUMMARY

[0004] The present application provides an image processing method, device, equipment and medium to solve the problem that the fixedly set feature acquisition model in the image processing scheme in the related art cannot process the surrounding environment images in different vehicle use scenarios, and cannot guarantee the accuracy of the bird's eye view feature extraction.

[0005] According to an aspect of the present application, an image processing method is provided, comprising:

[0006] obtaining a vehicle use scenario parameter corresponding to a to-be-processed surrounding environment image;

[0007] determining a target feature acquisition model corresponding to the vehicle use scenario of the to-be-processed surrounding environment image from the pre-trained various candidate feature acquisition models according to the vehicle use scenario parameter;

[0008] inputting the to-be-processed surrounding environment image into the target feature acquisition model to obtain an aerial view feature corresponding to the to-be-processed surrounding environment image;

[0009] The input of the target feature acquisition model is a surrounding environment image, and the output is an aerial view feature corresponding to the surrounding environment image.

[0010] According to another aspect of the present application, an image processing apparatus is provided, comprising:

[0011] a parameter acquisition module configured to acquire a vehicle use scenario parameter corresponding to a to-be-processed surrounding environment image;

[0012] a model determination module configured to determine, according to the vehicle use scenario parameter, a target feature acquisition model corresponding to a vehicle use scenario of the to-be-processed surrounding environment image from each candidate feature acquisition model pre-trained;

[0013] a feature acquisition module configured to input the to-be-processed surrounding environment image into the target feature acquisition model to obtain an aerial view feature corresponding to the to-be-processed surrounding environment image;

[0014] The input of the target feature acquisition model is a surrounding environment image, and the output is an aerial view feature corresponding to the surrounding environment image.

[0015] According to another aspect of the present application, an electronic device is provided, comprising:

[0016] at least one processor;

[0017] and a memory in network communication connection with the at least one processor;

[0018] The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor to enable the at least one processor to execute the image processing method according to any one of the embodiments of the present application.

[0019] According to another aspect of the present application, a computer readable storage medium is provided, which stores computer instructions for enabling a processor to execute the image processing method according to any one of the embodiments of the present application when executed.

[0020] The technical scheme of the embodiment of the present application comprises the following steps: obtaining a vehicle use scene parameter corresponding to a to-be-processed surrounding environment image; determining a target feature acquisition model corresponding to the vehicle use scene of the to-be-processed surrounding environment image from the pre-trained various candidate feature acquisition models according to the vehicle use scene parameter; and inputting the to-be-processed surrounding environment image into the target feature acquisition model to obtain the bird's eye view feature corresponding to the to-be-processed surrounding environment image. The problem that the fixedly set feature acquisition model in the image processing scheme in the related art cannot process the surrounding environment image under different vehicle use scenes and cannot guarantee the accuracy of the bird's eye view feature extraction is solved, and the beneficial effects that the feature acquisition model corresponding to the vehicle use scene of the surrounding environment image can be determined according to the vehicle use scene parameter corresponding to the surrounding environment image, the bird's eye view feature corresponding to the surrounding environment image can be quickly and accurately acquired based on the feature acquisition model corresponding to the vehicle use scene of the surrounding environment image, the surrounding environment image under different vehicle use scenes can be processed based on the candidate feature acquisition model corresponding to each vehicle use scene, and the bird's eye view feature corresponding to the surrounding environment image under different vehicle use scenes can be accurately acquired, and the accuracy of the bird's eye view feature acquisition is improved.

[0021] It should be understood that the content described in this part is not intended to identify key or important features of the embodiments of the present application, nor is it used to limit the scope of the present application. Other features of the present application will become apparent from the following description. BRIEF DESCRIPTION OF DRAWINGS

[0022] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0023] Figure 1 A flowchart of an image processing method provided for the first embodiment of the present application.

[0024] Figure 2 A flowchart of an image processing method provided for the second embodiment of the present application.

[0025] Figure 3 A flowchart of an image processing method provided for the third embodiment of the present application.

[0026] Figure 4 A structural schematic diagram of an image processing device provided for the fourth embodiment of the present application.

[0027] Figure 5 A structural schematic diagram of an electronic device for implementing the image processing method of the present application. DETAILED DESCRIPTION

[0028] In order to make the person skilled in the art better understand the present application, the technical solutions in the embodiments of the present application will be described clearly and completely in combination with the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by the person skilled in the art without creative labor should belong to the protection scope of the present application.

[0029] It should be noted that the terms "target", "first", "second" and the like in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and do not necessarily have to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances, so that the embodiments of the application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "include", "contain" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product or device containing a series of steps or units does not have to be limited to those steps or units clearly listed, but can include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0030] Embodiment one

[0031] Figure 1 A flow chart of an image processing method provided by the first embodiment of the present application. The present embodiment can be applied to the case where the bird's eye view feature corresponding to the surrounding environment image is obtained after the surrounding environment image of the vehicle is acquired. The method can be executed by an image processing device, which can be realized in the form of hardware and / or software, and can be configured in an electronic device. Illustratively, the electronic device can be an electronic control unit provided on the vehicle. As shown in the figure, the method comprises: Figure 1

[0032] Step 101, acquiring a vehicle use scene parameter corresponding to the surrounding environment image to be processed.

[0033] ​Optionally, the aerial view features corresponding to the surrounding environment image are image features of the surrounding environment image projected onto the grid map of the aerial view. The image features of the surrounding environment image can be features of each pixel point in the surrounding environment image. The features of each pixel point include but are not limited to the position and pixel value of each pixel point. The surrounding environment image to be processed is the surrounding environment image that needs to be processed and obtain corresponding aerial view features. The surrounding environment image to be processed includes images in the front left, front, front right, rear left, rear and rear right directions of the vehicle collected at the same time.

[0034] Optionally, the vehicle use scene parameter corresponding to the surrounding environment image to be processed is a parameter used to determine the vehicle use scene of the surrounding environment image to be processed. The vehicle use scene of the surrounding environment image to be processed is the use scene of the vehicle when the surrounding environment image to be processed is collected. The use scene of the vehicle includes but is not limited to: high-speed driving, medium-speed driving, low-speed driving, driving on the outermost lane of the road, driving on the innermost lane of the road, and driving on the middle lane of the road.

[0035] Optionally, the use scene of the vehicle is high-speed driving, indicating that the vehicle is driving at a high speed. The use scene of the vehicle is medium-speed driving, indicating that the vehicle is driving at a medium speed. The use scene of the vehicle is low-speed driving, indicating that the vehicle is driving at a low speed. The use scene of the vehicle is driving on the outermost lane of the road, indicating that the vehicle is driving on the outermost lane of the road. The use scene of the vehicle is driving on the innermost lane of the road, indicating that the vehicle is driving on the innermost lane of the road. The use scene of the vehicle is driving on the middle lane of the road, indicating that the vehicle is driving on the middle lane of the road. The middle lane of the road is the lane between the outermost lane and the innermost lane of the road.

[0036] Optionally, the vehicle use scene parameter corresponding to the surrounding environment image to be processed can be the vehicle speed corresponding to the surrounding environment image to be processed. The vehicle speed corresponding to the surrounding environment image to be processed is the vehicle speed of the vehicle when the surrounding environment image to be processed is collected. The electronic device can obtain the vehicle speed corresponding to the surrounding environment image to be processed from the vehicle speed sensor provided on the vehicle.

[0037] Optionally, generally, if the use scenario of the vehicle is high-speed driving, the vehicle speed is greater than or equal to the first vehicle speed threshold. If the use scenario of the vehicle is medium-speed driving, the vehicle speed is less than the first vehicle speed threshold and greater than or equal to the second vehicle speed threshold. If the use scenario of the vehicle is low-speed driving, the vehicle speed is less than the second vehicle speed threshold. Therefore, if the vehicle speed corresponding to the to-be-processed surrounding environment image is greater than or equal to the first vehicle speed threshold, it can be determined that the use scenario of the vehicle corresponding to the to-be-processed surrounding environment image is high-speed driving. If the vehicle speed corresponding to the to-be-processed surrounding environment image is less than the first vehicle speed threshold and greater than or equal to the second vehicle speed threshold, it can be determined that the use scenario of the vehicle corresponding to the to-be-processed surrounding environment image is medium-speed driving. If the vehicle speed corresponding to the to-be-processed surrounding environment image is less than the second vehicle speed threshold, it can be determined that the use scenario of the vehicle corresponding to the to-be-processed surrounding environment image is low-speed driving. The first vehicle speed threshold and the second vehicle speed threshold are two vehicle speed thresholds preset in advance. The first vehicle speed threshold is greater than the second vehicle speed threshold. For example, the first vehicle speed threshold is 90 kilometers per hour, and the second vehicle speed threshold is 60 kilometers per hour.

[0038] Optionally, the vehicle use scenario parameter corresponding to the to-be-processed surrounding environment image can be a vehicle position corresponding to the to-be-processed surrounding environment image. The vehicle position corresponding to the to-be-processed surrounding environment image is the position of the vehicle when the to-be-processed surrounding environment image is collected. The vehicle position corresponding to the to-be-processed surrounding environment image is the outermost lane, the innermost lane or the middle lane. The electronic device can obtain the vehicle position corresponding to the to-be-processed surrounding environment image from a positioning module arranged on the vehicle. The positioning module can be a global positioning system positioning module arranged on the vehicle.

[0039] Optionally, generally, if the use scenario of the vehicle is driving on the outermost lane of the road, the vehicle position is the outermost lane. If the use scenario of the vehicle is driving on the innermost lane of the road, the vehicle position is the innermost lane. If the use scenario of the vehicle is driving on the middle lane of the road, the vehicle position is the middle lane. Therefore, if the vehicle position corresponding to the to-be-processed surrounding environment image is the outermost lane, it can be determined that the use scenario of the vehicle corresponding to the to-be-processed surrounding environment image is driving on the outermost lane of the road. If the vehicle position corresponding to the to-be-processed surrounding environment image is the innermost lane, it can be determined that the use scenario of the vehicle corresponding to the to-be-processed surrounding environment image is driving on the innermost lane of the road. If the vehicle position corresponding to the to-be-processed surrounding environment image is the middle lane, it can be determined that the use scenario of the vehicle corresponding to the to-be-processed surrounding environment image is driving on the middle lane of the road.

[0040] Step 102, according to the vehicle use scenario parameter, determining the target feature acquisition model corresponding to the vehicle use scenario of the to-be-processed surrounding environment image from the pre-trained various candidate feature acquisition models.

[0041] Optionally, the pre-trained candidate feature acquisition model corresponding to each vehicle use scenario is a feature acquisition model trained according to surrounding environment images collected in the vehicle use scenario, so as to correspond to the bird's eye view features of the surrounding environment images collected in the vehicle use scenario. Therefore, the pre-trained candidate feature acquisition model corresponding to the vehicle use scenario can process the surrounding environment images in the vehicle use scenario and accurately acquire the bird's eye view features corresponding to the surrounding environment images in the vehicle use scenario.

[0042] Optionally, before acquiring the vehicle use scenario parameter corresponding to the to-be-processed surrounding environment image, the following operations are further included: for each vehicle use scenario, the following operations are performed: acquiring a preset number of surrounding environment images corresponding to the vehicle use scenario and bird's eye view features corresponding to each of the surrounding environment images; the vehicle use scenario is high-speed driving, medium-speed driving, or low-speed driving; and each of the surrounding environment images and the bird's eye view features corresponding to each of the surrounding environment images are used as training samples to train a machine learning model, so as to obtain a candidate feature acquisition model corresponding to the vehicle use scenario.

[0043] Optionally, a preset number of surrounding environment images corresponding to high-speed driving and bird's eye view features corresponding to each of the surrounding environment images are acquired. The preset number of surrounding environment images corresponding to high-speed driving are a plurality of surrounding environment images collected when the vehicle use scenario is high-speed driving. Each of the surrounding environment images is each surrounding environment image collected when the vehicle use scenario is high-speed driving. The bird's eye view features corresponding to each of the surrounding environment images are bird's eye view features corresponding to each surrounding environment image collected when the vehicle use scenario is high-speed driving. Each of the surrounding environment images and the bird's eye view features corresponding to each of the surrounding environment images are used as training samples to train a pre-set machine learning model, determine the parameters of the machine learning model, and obtain a candidate feature acquisition model corresponding to high-speed driving. The input of the candidate feature acquisition model is a surrounding environment image, and the output is a bird's eye view feature corresponding to the surrounding environment image. The candidate feature acquisition model corresponding to high-speed driving can process the surrounding environment images collected when the vehicle use scenario is high-speed driving, and accurately acquire the bird's eye view features corresponding to the surrounding environment images collected when the vehicle use scenario is high-speed driving.

[0044] Optionally, the preset number can be set according to business needs. For example, the preset number is 1000. 1000 surrounding environment images corresponding to high-speed driving and the bird's eye view features corresponding to each of the surrounding environment images are obtained.

[0045] Optionally, obtaining the preset number of surrounding environment images corresponding to high-speed driving and the bird's eye view features corresponding to each of the surrounding environment images includes: obtaining the preset number of surrounding environment images corresponding to high-speed driving and the bird's eye view features corresponding to each of the surrounding environment images uploaded by a target user through a terminal device. The target user uploads the preset number of surrounding environment images corresponding to high-speed driving and the bird's eye view features corresponding to each of the surrounding environment images to the target device through the terminal device. The target user can be a technical personnel responsible for managing the vehicle.

[0046] Optionally, the machine learning model includes but is not limited to a neural network model.

[0047] Optionally, training the preset machine learning model using each of the surrounding environment images and the bird's eye view features corresponding to each of the surrounding environment images as training samples to obtain the alternative feature acquisition model corresponding to high-speed driving includes: dividing the training samples composed of each of the surrounding environment images and the bird's eye view features corresponding to each of the surrounding environment images into a training sample set and a test sample set; training the machine learning model using the training sample set to obtain the alternative feature acquisition model corresponding to high-speed driving; and testing the alternative feature acquisition model corresponding to high-speed driving using the test sample set to obtain the accuracy of the alternative feature acquisition model corresponding to high-speed driving.

[0048] Optionally, the training samples composed of each of the surrounding environment images and the bird's eye view features corresponding to each of the surrounding environment images are divided into the training sample set and the test sample set according to a preset division ratio.

[0049] In one specific example, the training samples include 1000 surrounding environment images corresponding to high-speed driving and the bird's eye view features corresponding to each of the surrounding environment images. 80% of the sample data of the training samples, i.e., 800 surrounding environment images corresponding to high-speed driving and the bird's eye view features corresponding to each of the surrounding environment images, are divided into the training sample set. 20% of the sample data of the training samples, i.e., 200 surrounding environment images corresponding to high-speed driving and the bird's eye view features corresponding to each of the surrounding environment images, are divided into the test sample set.

[0050] Optionally, the test sample set is used to test the candidate feature acquisition model corresponding to high-speed driving, to obtain the accuracy of the candidate feature acquisition model corresponding to high-speed driving, including: inputting each surrounding environment image corresponding to high-speed driving in the test sample set into the candidate feature acquisition model corresponding to high-speed driving, to obtain the bird's eye view features corresponding to each surrounding environment image corresponding to high-speed driving output by the candidate feature acquisition model corresponding to high-speed driving; using a preset similarity algorithm to determine the similarity between the bird's eye view features corresponding to each surrounding environment image and the bird's eye view features corresponding to each surrounding environment image output by the candidate feature acquisition model; calculating the average value of the similarity between the bird's eye view features corresponding to each surrounding environment image and the bird's eye view features corresponding to each surrounding environment image output by the candidate feature acquisition model; and determining the average value as the accuracy of the candidate feature acquisition model corresponding to high-speed driving.

[0051] Optionally, when the accuracy of the candidate feature acquisition model corresponding to high-speed driving is greater than or equal to a preset accuracy threshold, it is determined that the candidate feature acquisition model corresponding to high-speed driving can be used. When the accuracy of the candidate feature acquisition model corresponding to high-speed driving is less than the preset accuracy threshold, the training of the candidate feature acquisition model corresponding to high-speed driving is continued. The preset accuracy threshold can be set according to business requirements. For example, the preset accuracy threshold is 0.9.

[0052] Optionally, a preset number of surrounding environment images corresponding to medium-speed driving and bird's eye view features corresponding to each of the surrounding environment images are obtained. The preset number of surrounding environment images corresponding to medium-speed driving are a plurality of surrounding environment images collected when the use scenario of the vehicle is medium-speed driving. Each of the surrounding environment images is a surrounding environment image collected when the use scenario of the vehicle is medium-speed driving. The bird's eye view features corresponding to each of the surrounding environment images are bird's eye view features corresponding to each of the surrounding environment images collected when the use scenario of the vehicle is medium-speed driving. Each of the surrounding environment images and the bird's eye view features corresponding to each of the surrounding environment images are used as training samples to train a pre-set machine learning model, to determine the parameters of the machine learning model, and to obtain a candidate feature acquisition model corresponding to medium-speed driving. The input of the candidate feature acquisition model is a surrounding environment image, and the output is a bird's eye view feature corresponding to the surrounding environment image. The candidate feature acquisition model corresponding to medium-speed driving can process the surrounding environment images collected when the use scenario of the vehicle is medium-speed driving, and accurately obtain the bird's eye view features corresponding to the surrounding environment images collected when the use scenario of the vehicle is medium-speed driving.

[0053] Optionally, the obtaining the preset number of surrounding environment images corresponding to the medium-speed driving and the bird's eye view features corresponding to each of the surrounding environment images comprises: obtaining the preset number of surrounding environment images corresponding to the medium-speed driving and the bird's eye view features corresponding to each of the surrounding environment images uploaded by a target user.

[0054] Optionally, the training the preset machine learning model using each of the surrounding environment images and the bird's eye view features corresponding to each of the surrounding environment images as a training sample to obtain the candidate feature acquisition model corresponding to the medium-speed driving comprises: dividing the training sample composed of each of the surrounding environment images and the bird's eye view features corresponding to each of the surrounding environment images into a training sample set and a test sample set; training the machine learning model using the training sample set to obtain the candidate feature acquisition model corresponding to the medium-speed driving; and testing the candidate feature acquisition model corresponding to the medium-speed driving using the test sample set to obtain the accuracy of the candidate feature acquisition model corresponding to the medium-speed driving.

[0055] Optionally, the testing the candidate feature acquisition model corresponding to the medium-speed driving using the test sample set to obtain the accuracy of the candidate feature acquisition model corresponding to the medium-speed driving comprises: inputting each of the surrounding environment images corresponding to the medium-speed driving in the test sample set into the candidate feature acquisition model corresponding to the medium-speed driving to obtain the bird's eye view features corresponding to each of the surrounding environment images corresponding to the medium-speed driving output by the candidate feature acquisition model corresponding to the medium-speed driving; determining the similarity between the bird's eye view features corresponding to each of the surrounding environment images and the bird's eye view features corresponding to each of the surrounding environment images output by the candidate feature acquisition model corresponding to the medium-speed driving using a preset similarity algorithm; calculating the average value of the similarity between the bird's eye view features corresponding to each of the surrounding environment images and the bird's eye view features corresponding to each of the surrounding environment images output by the candidate feature acquisition model corresponding to the medium-speed driving; and determining the average value as the accuracy of the candidate feature acquisition model corresponding to the medium-speed driving.

[0056] Optionally, when the accuracy of the candidate feature acquisition model corresponding to the medium-speed driving is greater than or equal to a preset accuracy threshold, it is determined that the candidate feature acquisition model corresponding to the medium-speed driving can be used. When the accuracy of the candidate feature acquisition model corresponding to the medium-speed driving is less than the preset accuracy threshold, the training of the candidate feature acquisition model corresponding to the medium-speed driving is continued.

[0057] Optionally, a preset number of surrounding environment images corresponding to low-speed driving and bird's eye view features corresponding to each of the surrounding environment images are acquired. The preset number of surrounding environment images corresponding to low-speed driving are a plurality of surrounding environment images collected when the use scenario of the vehicle is low-speed driving. Each of the surrounding environment images is a respective surrounding environment image collected when the use scenario of the vehicle is low-speed driving. The bird's eye view features corresponding to each of the surrounding environment images are bird's eye view features corresponding to each of the surrounding environment images collected when the use scenario of the vehicle is low-speed driving. Each of the surrounding environment images and the bird's eye view features corresponding to each of the surrounding environment images are used as training samples to train a pre-configured machine learning model, to determine parameters of the machine learning model, and to obtain an alternative feature acquisition model corresponding to low-speed driving. The input of the alternative feature acquisition model corresponding to low-speed driving is a surrounding environment image, and the output is a bird's eye view feature corresponding to the surrounding environment image. The alternative feature acquisition model corresponding to low-speed driving can process the surrounding environment images collected when the use scenario of the vehicle is low-speed driving, and accurately acquire the bird's eye view features corresponding to the surrounding environment images collected when the use scenario of the vehicle is low-speed driving.

[0058] Optionally, acquiring a preset number of surrounding environment images corresponding to low-speed driving and bird's eye view features corresponding to each of the surrounding environment images includes: acquiring a preset number of surrounding environment images corresponding to low-speed driving and bird's eye view features corresponding to each of the surrounding environment images uploaded by a target user to a target device through a terminal device.

[0059] Optionally, using each of the surrounding environment images and the bird's eye view features corresponding to each of the surrounding environment images as training samples to train a pre-configured machine learning model to obtain an alternative feature acquisition model corresponding to low-speed driving includes: dividing training samples composed of each of the surrounding environment images and the bird's eye view features corresponding to each of the surrounding environment images into a training sample set and a test sample set; using the training sample set to train the machine learning model to obtain the alternative feature acquisition model corresponding to low-speed driving; and using the test sample set to test the alternative feature acquisition model corresponding to low-speed driving to obtain the accuracy of the alternative feature acquisition model corresponding to low-speed driving.

[0060] Optionally, the accuracy of the candidate feature acquisition model corresponding to low-speed driving is obtained by testing the candidate feature acquisition model corresponding to low-speed driving using the test sample set, comprising: inputting each surrounding environment image corresponding to low-speed driving in the test sample set into the candidate feature acquisition model corresponding to low-speed driving to obtain the bird's eye view feature corresponding to each surrounding environment image corresponding to low-speed driving output by the candidate feature acquisition model corresponding to low-speed driving; determining the similarity between the bird's eye view feature corresponding to each surrounding environment image and the bird's eye view feature corresponding to each surrounding environment image output by the candidate feature acquisition model using a preset similarity algorithm; calculating the average value of the similarity between the bird's eye view feature corresponding to each surrounding environment image and the bird's eye view feature corresponding to each surrounding environment image output by the candidate feature acquisition model; and determining the average value as the accuracy of the candidate feature acquisition model corresponding to low-speed driving.

[0061] Optionally, when the accuracy of the candidate feature acquisition model corresponding to low-speed driving is greater than or equal to a preset accuracy threshold, it is determined that the candidate feature acquisition model corresponding to low-speed driving can be used. When the accuracy of the candidate feature acquisition model corresponding to low-speed driving is less than the preset accuracy threshold, the training of the candidate feature acquisition model corresponding to low-speed driving is continued.

[0062] Optionally, the target feature acquisition model corresponding to the vehicle use scenario of the surrounding environment image to be processed is a feature acquisition model that can process the surrounding environment image in the vehicle use scenario of the surrounding environment image to be processed and accurately acquire the bird's eye view feature corresponding to the surrounding environment image in the vehicle use scenario of the surrounding environment image to be processed.

[0063] Optionally, the vehicle use scenario parameter is vehicle speed; and the target feature acquisition model corresponding to the vehicle use scenario of the surrounding environment image to be processed is determined from the pre-trained candidate feature acquisition models according to the vehicle use scenario parameter, comprising: determining whether the vehicle speed corresponding to the surrounding environment image to be processed is greater than or equal to a first vehicle speed threshold; and if the vehicle speed corresponding to the surrounding environment image to be processed is greater than or equal to the first vehicle speed threshold, determining that the vehicle use scenario of the surrounding environment image to be processed is high-speed driving and determining the candidate feature acquisition model corresponding to high-speed driving as the target feature acquisition model corresponding to the vehicle use scenario of the surrounding environment image to be processed.

[0064] Optionally, if the vehicle usage scenario of the to-be-processed surrounding environment image is high-speed driving, the alternative feature acquisition model corresponding to high-speed driving is a feature acquisition model that can process the surrounding environment image in the vehicle usage scenario of the to-be-processed surrounding environment image and accurately acquire the bird's eye view feature corresponding to the surrounding environment image in the vehicle usage scenario of the to-be-processed surrounding environment image. Therefore, after determining that the vehicle usage scenario of the to-be-processed surrounding environment image is high-speed driving according to the vehicle speed corresponding to the to-be-processed surrounding environment image, the alternative feature acquisition model corresponding to high-speed driving is determined as the target feature acquisition model corresponding to the vehicle usage scenario of the to-be-processed surrounding environment image.

[0065] Optionally, after determining whether the vehicle speed corresponding to the to-be-processed surrounding environment image is greater than or equal to the first vehicle speed threshold, the method further includes: if the vehicle speed corresponding to the to-be-processed surrounding environment image is less than the first vehicle speed threshold, determining whether the vehicle speed corresponding to the to-be-processed surrounding environment image is greater than or equal to a second vehicle speed threshold; and if the vehicle speed corresponding to the to-be-processed surrounding environment image is greater than or equal to the second vehicle speed threshold, determining that the vehicle usage scenario of the to-be-processed surrounding environment image is medium-speed driving, and determining the alternative feature acquisition model corresponding to medium-speed driving as the target feature acquisition model corresponding to the vehicle usage scenario of the to-be-processed surrounding environment image.

[0066] Optionally, if the vehicle usage scenario of the to-be-processed surrounding environment image is medium-speed driving, the alternative feature acquisition model corresponding to medium-speed driving is a feature acquisition model that can process the surrounding environment image in the vehicle usage scenario of the to-be-processed surrounding environment image and accurately acquire the bird's eye view feature corresponding to the surrounding environment image in the vehicle usage scenario of the to-be-processed surrounding environment image. Therefore, after determining that the vehicle usage scenario of the to-be-processed surrounding environment image is medium-speed driving according to the vehicle speed corresponding to the to-be-processed surrounding environment image, the alternative feature acquisition model corresponding to medium-speed driving is determined as the target feature acquisition model corresponding to the vehicle usage scenario of the to-be-processed surrounding environment image.

[0067] Optionally, after determining whether the vehicle speed corresponding to the to-be-processed surrounding environment image is greater than or equal to the second vehicle speed threshold, the method further includes: if the vehicle speed corresponding to the to-be-processed surrounding environment image is less than the second vehicle speed threshold, determining that the vehicle usage scenario of the to-be-processed surrounding environment image is low-speed driving, and determining the alternative feature acquisition model corresponding to low-speed driving as the target feature acquisition model corresponding to the vehicle usage scenario of the to-be-processed surrounding environment image.

[0068] Optionally, if the vehicle usage scenario of the to-be-processed surrounding environment image is low-speed driving, the alternative feature acquisition model corresponding to low-speed driving is a feature acquisition model that can process the surrounding environment image in the vehicle usage scenario of the to-be-processed surrounding environment image and accurately acquire the bird's eye view feature corresponding to the surrounding environment image in the vehicle usage scenario of the to-be-processed surrounding environment image. Therefore, after determining that the vehicle usage scenario of the to-be-processed surrounding environment image is low-speed driving according to the vehicle speed corresponding to the to-be-processed surrounding environment image, the alternative feature acquisition model corresponding to low-speed driving is determined as the target feature acquisition model corresponding to the vehicle usage scenario of the to-be-processed surrounding environment image.

[0069] Optionally, before acquiring the vehicle usage scenario parameter corresponding to the to-be-processed surrounding environment image, the following operations are further included: for each vehicle usage scenario, the following operations are performed: acquiring a preset number of surrounding environment images corresponding to the vehicle usage scenario and bird's eye view features corresponding to each of the surrounding environment images; wherein the vehicle usage scenario is driving on the outermost lane of the road, driving on the innermost lane of the road, or driving on the middle lane of the road; using each of the surrounding environment images and the bird's eye view features corresponding to each of the surrounding environment images as training samples, training a machine learning model to obtain an alternative feature extraction model corresponding to the vehicle usage scenario.

[0070] Optionally, a preset number of surrounding environment images corresponding to driving on the outermost lane of the road and bird's eye view features corresponding to each of the surrounding environment images are acquired. The preset number of surrounding environment images corresponding to driving on the outermost lane of the road are a plurality of surrounding environment images collected when the vehicle usage scenario is driving on the outermost lane of the road. Each of the surrounding environment images is each surrounding environment image collected when the vehicle usage scenario is driving on the outermost lane of the road. The bird's eye view features corresponding to each of the surrounding environment images are bird's eye view features corresponding to each surrounding environment image collected when the vehicle usage scenario is driving on the outermost lane of the road. Each of the surrounding environment images and the bird's eye view features corresponding to each of the surrounding environment images are used as training samples to train a pre-set machine learning model, determine parameters of the machine learning model, and obtain an alternative feature acquisition model corresponding to driving on the outermost lane of the road. The input of the alternative feature acquisition model is a surrounding environment image, and the output is a bird's eye view feature corresponding to the surrounding environment image. The alternative feature acquisition model corresponding to driving on the outermost lane of the road can process the surrounding environment image collected when the vehicle usage scenario is driving on the outermost lane of the road and accurately acquire the bird's eye view feature corresponding to the surrounding environment image collected when the vehicle usage scenario is driving on the outermost lane of the road.

[0071] Optionally, a preset number of surrounding environment images corresponding to driving in the innermost lane of the road and aerial view features corresponding to each of the surrounding environment images are acquired. The preset number of surrounding environment images corresponding to driving in the innermost lane of the road are a plurality of surrounding environment images collected when the use scenario of the vehicle is driving in the innermost lane of the road. Each of the surrounding environment images is each surrounding environment image collected when the use scenario of the vehicle is driving in the innermost lane of the road. The aerial view features corresponding to each of the surrounding environment images are aerial view features corresponding to each of the surrounding environment images collected when the use scenario of the vehicle is driving in the innermost lane of the road. Each of the surrounding environment images and the aerial view features corresponding to each of the surrounding environment images are taken as training samples, a pre-set machine learning model is trained, parameters of the machine learning model are determined, and an alternative feature acquisition model corresponding to driving in the innermost lane of the road is obtained. The input of the alternative feature acquisition model is a surrounding environment image, and the output is an aerial view feature corresponding to the surrounding environment image. The alternative feature acquisition model corresponding to driving in the innermost lane of the road can process the surrounding environment image collected when the use scenario of the vehicle is driving in the innermost lane of the road, and accurately acquire the aerial view feature corresponding to the surrounding environment image collected when the use scenario of the vehicle is driving in the innermost lane of the road.

[0072] Optionally, a preset number of surrounding environment images corresponding to driving in the middle lane of the road and aerial view features corresponding to each of the surrounding environment images are acquired. The preset number of surrounding environment images corresponding to driving in the middle lane of the road are a plurality of surrounding environment images collected when the use scenario of the vehicle is driving in the middle lane of the road. Each of the surrounding environment images is each surrounding environment image collected when the use scenario of the vehicle is driving in the middle lane of the road. The aerial view features corresponding to each of the surrounding environment images are aerial view features corresponding to each of the surrounding environment images collected when the use scenario of the vehicle is driving in the middle lane of the road. Each of the surrounding environment images and the aerial view features corresponding to each of the surrounding environment images are taken as training samples, a pre-set machine learning model is trained, parameters of the machine learning model are determined, and an alternative feature acquisition model corresponding to driving in the middle lane of the road is obtained. The input of the alternative feature acquisition model is a surrounding environment image, and the output is an aerial view feature corresponding to the surrounding environment image. The alternative feature acquisition model corresponding to driving in the middle lane of the road can process the surrounding environment image collected when the use scenario of the vehicle is driving in the middle lane of the road, and accurately acquire the aerial view feature corresponding to the surrounding environment image collected when the use scenario of the vehicle is driving in the middle lane of the road.

[0073] Optionally, the vehicle use scene parameter is a vehicle position; and according to the vehicle use scene parameter, the target feature extraction model corresponding to the vehicle use scene of the to-be-processed surrounding environment image is determined from the pre-trained candidate feature extraction models, including: if the vehicle position corresponding to the to-be-processed surrounding environment image is an outermost lane, it is determined that the vehicle use scene of the to-be-processed surrounding environment image is driving on the outermost lane of the road, and the candidate feature extraction model corresponding to driving on the outermost lane of the road is determined as the target feature extraction model corresponding to the vehicle use scene of the to-be-processed surrounding environment image; if the vehicle position corresponding to the to-be-processed surrounding environment image is an innermost lane, it is determined that the vehicle use scene of the to-be-processed surrounding environment image is driving on the innermost lane of the road, and the candidate feature extraction model corresponding to driving on the innermost lane of the road is determined as the target feature extraction model corresponding to the vehicle use scene of the to-be-processed surrounding environment image; if the vehicle position corresponding to the to-be-processed surrounding environment image is a middle lane, it is determined that the vehicle use scene of the to-be-processed surrounding environment image is driving on the middle lane of the road, and the candidate feature extraction model corresponding to driving on the middle lane of the road is determined as the target feature extraction model corresponding to the vehicle use scene of the to-be-processed surrounding environment image.

[0074] Optionally, if the vehicle use scene of the to-be-processed surrounding environment image is driving on the outermost lane of the road, the candidate feature extraction model corresponding to driving on the outermost lane of the road is the feature extraction model that can process the surrounding environment image in the vehicle use scene of the to-be-processed surrounding environment image and accurately extract the bird's eye view feature corresponding to the surrounding environment image in the vehicle use scene of the to-be-processed surrounding environment image. Therefore, after it is determined that the vehicle use scene of the to-be-processed surrounding environment image is driving on the outermost lane of the road according to the vehicle speed corresponding to the to-be-processed surrounding environment image, the candidate feature extraction model corresponding to driving on the outermost lane of the road is determined as the target feature extraction model corresponding to the vehicle use scene of the to-be-processed surrounding environment image.

[0075] Optionally, if the vehicle use scene of the to-be-processed surrounding environment image is driving on the innermost lane of the road, the candidate feature extraction model corresponding to driving on the innermost lane of the road is the feature extraction model that can process the surrounding environment image in the vehicle use scene of the to-be-processed surrounding environment image and accurately extract the bird's eye view feature corresponding to the surrounding environment image in the vehicle use scene of the to-be-processed surrounding environment image. Therefore, after it is determined that the vehicle use scene of the to-be-processed surrounding environment image is driving on the innermost lane of the road according to the vehicle speed corresponding to the to-be-processed surrounding environment image, the candidate feature extraction model corresponding to driving on the innermost lane of the road is determined as the target feature extraction model corresponding to the vehicle use scene of the to-be-processed surrounding environment image.

[0076] Optionally, if the vehicle usage scenario of the to-be-processed surrounding environment image is driving in the middle lane of the road, the alternative feature acquisition model corresponding to driving in the middle lane of the road is a feature acquisition model that can process the surrounding environment image in the vehicle usage scenario of the to-be-processed surrounding environment image and accurately acquire the bird's eye view feature corresponding to the surrounding environment image in the vehicle usage scenario of the to-be-processed surrounding environment image. Therefore, after determining that the vehicle usage scenario of the to-be-processed surrounding environment image is driving in the middle lane of the road according to the vehicle speed corresponding to the to-be-processed surrounding environment image, the alternative feature acquisition model corresponding to driving in the middle lane of the road is determined as the target feature acquisition model corresponding to the vehicle usage scenario of the to-be-processed surrounding environment image.

[0077] Step 103, inputting the to-be-processed surrounding environment image into the target feature acquisition model to obtain the bird's eye view feature corresponding to the to-be-processed surrounding environment image.

[0078] The input of the target feature acquisition model is a surrounding environment image, and the output is a bird's eye view feature corresponding to the surrounding environment image.

[0079] Optionally, the to-be-processed surrounding environment image is input into the target feature acquisition model to obtain the bird's eye view feature corresponding to the to-be-processed surrounding environment image. Thus, according to the vehicle usage scenario parameter corresponding to the to-be-processed surrounding environment image, the feature acquisition model corresponding to the vehicle usage scenario of the to-be-processed surrounding environment image is determined, and based on the feature acquisition model corresponding to the vehicle usage scenario of the to-be-processed surrounding environment image, the bird's eye view feature corresponding to the to-be-processed surrounding environment image is quickly and accurately acquired.

[0080] The technical scheme of the embodiment of the present application acquires the vehicle use scene parameter corresponding to the to-be-processed surrounding environment image, then determines the target feature acquisition model corresponding to the vehicle use scene of the to-be-processed surrounding environment image from the pre-trained various candidate feature acquisition models according to the vehicle use scene parameter, and finally inputs the to-be-processed surrounding environment image into the target feature acquisition model to obtain the bird's eye view feature corresponding to the to-be-processed surrounding environment image, thereby solving the problem in the related art that the fixedly set feature acquisition model in the image processing scheme cannot process the surrounding environment image under different vehicle use scenes and cannot guarantee the accuracy of the bird's eye view feature extraction, and achieving the beneficial effects that the feature acquisition model corresponding to the vehicle use scene of the surrounding environment image can be determined according to the vehicle use scene parameter corresponding to the surrounding environment image, the bird's eye view feature corresponding to the surrounding environment image can be quickly and accurately acquired based on the feature acquisition model corresponding to the vehicle use scene of the surrounding environment image, the surrounding environment image under different vehicle use scenes can be processed based on the candidate feature acquisition model corresponding to each vehicle use scene, the bird's eye view feature corresponding to the surrounding environment image under different vehicle use scenes can be accurately acquired, and the accuracy of the bird's eye view feature acquisition is improved.

[0081] Embodiment Two

[0082] Figure 2 A flowchart of an image processing method provided by the embodiment two of the present application. The embodiment of the present application can be combined with each optional scheme in one or more of the above embodiments. As shown in the figure, the method comprises the following steps. Figure 2

[0083] Step 201: Acquire the vehicle speed corresponding to the to-be-processed surrounding environment image.

[0084] Step 202: Determine whether the vehicle speed corresponding to the to-be-processed surrounding environment image is greater than or equal to a first vehicle speed threshold value: if yes, execute step 203; if no, execute step 204.

[0085] Step 203: Determine that the vehicle use scene of the to-be-processed surrounding environment image is high-speed driving, and determine the candidate feature acquisition model corresponding to the high-speed driving as the target feature acquisition model corresponding to the vehicle use scene of the to-be-processed surrounding environment image.

[0086] Step 204: Determine whether the vehicle speed corresponding to the to-be-processed surrounding environment image is greater than or equal to a second vehicle speed threshold value: if yes, execute step 205; if no, execute step 206.

[0087] ​Step 205, determining that the vehicle use scene of the to-be-processed surrounding environment image is medium-speed driving, and determining the candidate feature acquisition model corresponding to the medium-speed driving as the target feature acquisition model corresponding to the vehicle use scene of the to-be-processed surrounding environment image.

[0088] Step 206, determining that the vehicle use scene of the to-be-processed surrounding environment image is low-speed driving, and determining the candidate feature acquisition model corresponding to the low-speed driving as the target feature acquisition model corresponding to the vehicle use scene of the to-be-processed surrounding environment image.

[0089] Step 207, inputting the to-be-processed surrounding environment image into the target feature acquisition model to obtain the bird's eye view feature corresponding to the to-be-processed surrounding environment image.

[0090] The input of the target feature acquisition model is a surrounding environment image, and the output is a bird's eye view feature corresponding to the surrounding environment image.

[0091] The technical scheme of the embodiment of the application can determine the feature acquisition model corresponding to the vehicle use scene of the surrounding environment image according to the vehicle speed corresponding to the surrounding environment image, can quickly and accurately acquire the bird's eye view feature corresponding to the surrounding environment image based on the feature acquisition model corresponding to the vehicle use scene of the surrounding environment image, and can accurately acquire the bird's eye view feature corresponding to the surrounding environment image under different vehicle use scenes based on the candidate feature acquisition models corresponding to the vehicle use scenes, thereby improving the accuracy of bird's eye view feature acquisition.

[0092] Embodiment three

[0093] Figure 3 A flowchart of an image processing method provided by the third embodiment of the application. The embodiment of the application can be combined with one or more optional schemes in the above embodiments. As shown in the figure, the method comprises the following steps. Figure 3

[0094] Step 301, acquiring the vehicle position corresponding to the to-be-processed surrounding environment image: if the vehicle position corresponding to the to-be-processed surrounding environment image is the outermost lane, performing step 302; if the vehicle position corresponding to the to-be-processed surrounding environment image is the innermost lane, performing step 303; if the vehicle position corresponding to the to-be-processed surrounding environment image is the middle lane, performing step 304.

[0095] Step 302, determining that the vehicle use scene of the to-be-processed surrounding environment image is driving on the outermost lane of the road, and determining the candidate feature acquisition model corresponding to driving on the outermost lane of the road as the target feature acquisition model corresponding to the vehicle use scene of the to-be-processed surrounding environment image.​

[0096] Step 303, determining that the vehicle use scene of the to-be-processed surrounding environment image is driving in the innermost lane of the road, and determining the candidate feature acquisition model corresponding to driving in the innermost lane of the road as the target feature acquisition model corresponding to the vehicle use scene of the to-be-processed surrounding environment image.

[0097] Step 304, determining that the vehicle use scene of the to-be-processed surrounding environment image is driving in the middle lane of the road, and determining the candidate feature acquisition model corresponding to driving in the middle lane of the road as the target feature acquisition model corresponding to the vehicle use scene of the to-be-processed surrounding environment image.

[0098] Step 305, inputting the to-be-processed surrounding environment image into the target feature acquisition model to obtain the bird's eye view feature corresponding to the to-be-processed surrounding environment image.

[0099] The input of the target feature acquisition model is a surrounding environment image, and the output is a bird's eye view feature corresponding to the surrounding environment image.

[0100] The technical scheme of the embodiment of the present application can determine the feature acquisition model corresponding to the vehicle use scene of the surrounding environment image according to the vehicle position corresponding to the surrounding environment image, can quickly and accurately acquire the bird's eye view feature corresponding to the surrounding environment image based on the feature acquisition model corresponding to the vehicle use scene of the surrounding environment image, and can accurately acquire the bird's eye view feature corresponding to the surrounding environment image in different vehicle use scenes based on the candidate feature acquisition model corresponding to each vehicle use scene, thereby improving the accuracy of bird's eye view feature acquisition.

[0101] Embodiment four

[0102] Figure 4 A structural schematic diagram of an image processing device provided by the fourth embodiment of the present application is shown. The device can be configured in an electronic device. As shown in the figure, the device comprises a parameter acquisition module 401, a model determination module 402, and a feature acquisition module 403. Figure 4

[0103] ​The parameter acquisition module 401 is configured to acquire a vehicle use scene parameter corresponding to a to-be-processed surrounding environment image; the model determination module 402 is configured to determine, according to the vehicle use scene parameter, a target feature acquisition model corresponding to a vehicle use scene of the to-be-processed surrounding environment image from each candidate feature acquisition model that is pre-trained; and the feature acquisition module 403 is configured to input the to-be-processed surrounding environment image into the target feature acquisition model to obtain an aerial view feature corresponding to the to-be-processed surrounding environment image; wherein the input of the target feature acquisition model is a surrounding environment image, and the output is an aerial view feature corresponding to the surrounding environment image.

[0104] The technical scheme of the embodiment of the application acquires a vehicle use scene parameter corresponding to a to-be-processed surrounding environment image, then determines, according to the vehicle use scene parameter, a target feature acquisition model corresponding to a vehicle use scene of the to-be-processed surrounding environment image from each candidate feature acquisition model that is pre-trained, finally inputs the to-be-processed surrounding environment image into the target feature acquisition model to obtain an aerial view feature corresponding to the to-be-processed surrounding environment image, thereby solving the problem in the related art that a fixedly set feature acquisition model cannot process surrounding environment images under different vehicle use scenes and cannot guarantee the accuracy of aerial view feature extraction, and achieving the beneficial effects that the feature acquisition model corresponding to the vehicle use scene of the surrounding environment image can be determined according to the vehicle use scene parameter corresponding to the surrounding environment image, the aerial view feature corresponding to the surrounding environment image can be quickly and accurately acquired based on the feature acquisition model corresponding to the vehicle use scene of the surrounding environment image, and the aerial view feature corresponding to the surrounding environment image under different vehicle use scenes can be accurately acquired based on the candidate feature acquisition model corresponding to each vehicle use scene, thereby improving the accuracy of aerial view feature acquisition.

[0105] In an optional implementation of the embodiment of the application, optionally, the vehicle use scene parameter is a vehicle speed; the model determination module 402 is specifically configured to: determine whether the vehicle speed corresponding to the to-be-processed surrounding environment image is greater than or equal to a first vehicle speed threshold; if the vehicle speed corresponding to the to-be-processed surrounding environment image is greater than or equal to the first vehicle speed threshold, it is determined that the vehicle use scene of the to-be-processed surrounding environment image is high-speed driving, and the candidate feature acquisition model corresponding to high-speed driving is determined as the target feature acquisition model corresponding to the vehicle use scene of the to-be-processed surrounding environment image.

[0106] In an optional implementation of the embodiment of the present application, optionally, the model determining module 402 is further specifically configured to: if the vehicle speed corresponding to the to-be-processed surrounding environment image is less than the first vehicle speed threshold, determine whether the vehicle speed corresponding to the to-be-processed surrounding environment image is greater than or equal to a second vehicle speed threshold; and if the vehicle speed corresponding to the to-be-processed surrounding environment image is greater than or equal to the second vehicle speed threshold, determine that the vehicle use scenario of the to-be-processed surrounding environment image is medium-speed driving, and determine the candidate feature acquisition model corresponding to the medium-speed driving as the target feature acquisition model corresponding to the vehicle use scenario of the to-be-processed surrounding environment image.

[0107] In an optional implementation of the embodiment of the present application, optionally, the model determining module 402 is further specifically configured to: if the vehicle speed corresponding to the to-be-processed surrounding environment image is less than the second vehicle speed threshold, determine that the vehicle use scenario of the to-be-processed surrounding environment image is low-speed driving, and determine the candidate feature acquisition model corresponding to the low-speed driving as the target feature acquisition model corresponding to the vehicle use scenario of the to-be-processed surrounding environment image.

[0108] In an optional implementation of the embodiment of the present application, optionally, the image processing apparatus further comprises a first model training module configured to: for each vehicle use scenario, perform the following operations: acquire a preset number of surrounding environment images corresponding to the vehicle use scenario and bird's eye view features corresponding to each of the surrounding environment images; wherein the vehicle use scenario is high-speed driving, medium-speed driving or low-speed driving; and use each of the surrounding environment images and the bird's eye view features corresponding to each of the surrounding environment images as a training sample to train a machine learning model, to obtain a candidate feature acquisition model corresponding to the vehicle use scenario.

[0109] In an optional implementation of the embodiment of the present application, optionally, the vehicle use scenario parameter is a vehicle position; the model determination module 402 is specifically configured to: if the vehicle position corresponding to the to-be-processed surrounding environment image is an outermost lane, determine that the vehicle use scenario of the to-be-processed surrounding environment image is driving on the outermost lane of the road, and determine the candidate feature acquisition model corresponding to driving on the outermost lane of the road as the target feature acquisition model corresponding to the vehicle use scenario of the to-be-processed surrounding environment image; if the vehicle position corresponding to the to-be-processed surrounding environment image is an innermost lane, determine that the vehicle use scenario of the to-be-processed surrounding environment image is driving on the innermost lane of the road, and determine the candidate feature acquisition model corresponding to driving on the innermost lane of the road as the target feature acquisition model corresponding to the vehicle use scenario of the to-be-processed surrounding environment image; if the vehicle position corresponding to the to-be-processed surrounding environment image is a middle lane, determine that the vehicle use scenario of the to-be-processed surrounding environment image is driving on the middle lane of the road, and determine the candidate feature acquisition model corresponding to driving on the middle lane of the road as the target feature acquisition model corresponding to the vehicle use scenario of the to-be-processed surrounding environment image.

[0110] In an optional implementation of the embodiment of the present application, optionally, the image processing apparatus further includes a second model training module configured to perform the following operations for each vehicle use scenario: obtaining a preset number of surrounding environment images corresponding to the vehicle use scenario and bird's eye view features corresponding to each of the surrounding environment images; wherein the vehicle use scenario is driving on the outermost lane of the road, driving on the innermost lane of the road, or driving on the middle lane of the road; and using each of the surrounding environment images and the bird's eye view features corresponding to each of the surrounding environment images as training samples to train a machine learning model to obtain a candidate feature acquisition model corresponding to the vehicle use scenario.

[0111] As to the apparatus in the above embodiments, the specific modes in which various modules perform operations have been described in details in the embodiments of the method, and will not be described in details here.

[0112] The image processing apparatus described above can perform the image processing method provided in any embodiment of the present application, and has the corresponding function modules and beneficial effects of performing the image processing method.

[0113] Embodiment five

[0114] Figure 5 A structure schematic diagram of an electronic device 10 that can be used to implement the image processing method of the embodiment of the present application is shown. The components shown herein, their connections and relationships, and their functions are merely examples, and are not intended to limit the implementation of the present application described and / or claimed herein.

[0115] As shown in Figure 5 The electronic device 10 includes at least one processor 11, and a memory, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc., in network communication connection with the at least one processor 11, where the memory stores computer programs executable by the at least one processor. The processor 11 can perform various appropriate actions and processes according to the computer programs stored in the read-only memory (ROM) 12 or the computer programs built from the storage unit 18 to the random access memory (RAM) 13. Various programs and data required for the operation of the electronic device 10 can also be stored in the random access memory (RAM) 13. The processor 11, the read-only memory (ROM) 12, and the random access memory (RAM) 13 are connected to each other through a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0116] Various components in the electronic device 10 are connected to the input / output (I / O) interface 15, including an input unit 16, such as a keyboard, a mouse, etc., an output unit 17, such as various types of displays, a speaker, etc., a storage unit 18, such as a magnetic disk, an optical disk, etc., and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices through a computer network, such as the Internet, and / or various telecommunication networks.

[0117] The processor 11 can be various general and / or special-purpose processing components with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 11 performs various methods and processes described above, such as the image processing method.

[0118] In some embodiments, the image processing method can be implemented as a computer program tangibly embodied in a computer readable storage medium, such as the storage unit 18. In some embodiments, part or all of the computer program can be loaded and / or installed onto the electronic device 10 via the read-only memory (ROM) 12 and / or the communication unit 19. When the computer program is built into the random access memory (RAM) 13 and executed by the processor 11, one or more steps of the image processing method described above can be performed. Alternatively, in other embodiments, the processor 11 can be configured to perform the image processing method by any other appropriate means, such as by means of firmware.

[0119] The various embodiments of the systems and techniques described above can be implemented in digital electronic circuitry, integrated circuitry, a field programmable gate array (FPGA), an application specific integrated circuit (ASIC), a system on a chip (SOC), a load programmable logic device (CPLD), computer hardware, firmware, software, and / or combinations thereof. These various embodiments can include implementation in one or more computer programs that are executable and / or interpretable on a programmable system including at least one programmable processor, which can be special or general purpose, coupled to receive data and instructions from, and to transmit data and instructions to, a storage system, at least one input device, and at least one output device.

[0120] Computer programs used to implement the image processing methods of the present application can be written in any combination of one or more programming languages. These computer programs can be provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine, such that the computer program, when executed, implements the functions / acts specified in the flowcharts and / or block diagrams. The computer program can be executed entirely on a machine, partially on a machine, partially on a machine as part of a standalone software package, or entirely on a remote machine or server.

[0121] In the context of the present application, a computer-readable storage medium can be a tangible medium that can contain or store computer programs for use by or in connection with an instruction execution system, apparatus, or device. Computer-readable storage media can include, but are not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium can be a machine-readable signal medium. More specific examples of the machine-readable storage medium will include one or more lines of electrical connections, portable computer disks, hard disk drives, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or Flash memory), optical fibers, portable compact disc read-only memories (CD-ROMs), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0122] To provide for interaction with a user, the systems and techniques described here can be implemented on an electronic device having a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user and a keyboard and a pointing device (e.g., a mouse or a trackball) by which the user can provide input to the electronic device. Other kinds of devices can be used to provide for interaction with a user as well; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form, including acoustic, speech, or tactile input.

[0123] The systems and techniques described here can be implemented in a computing system that includes a back end component (e.g., as a data server), or that includes a middleware component (e.g., an application server), or that includes a front end component (e.g., a user computer having a graphical user interface or a Web browser through which a user can interact with an implementation of the systems and techniques described here), or any combination of such back end, middleware, or front end components. The components of the system can be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0124] The computing system can include clients and servers. A client and server are generally remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other. A server can be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system, to solve the defects of large management difficulty and weak business scalability in traditional physical host and VPS service.

[0125] It should be understood that the various forms of flow shown above can be re-ordered, added to, or deleted from without departing from the scope of the present disclosure. For example, the steps recited in the present disclosure can be executed in parallel, executed in sequence, or executed in a different order, as long as the desired results of the present disclosure are achieved, and the present disclosure is not limited herein.

[0126] The specific embodiments described above are not intended to be limiting, and persons skilled in the art will appreciate that various modifications, combinations, sub-combinations and alternatives can be made to the specific embodiments without departing from the spirit and principles of the disclosure. Accordingly, the disclosure is not limited to the specific embodiments described above, but only by the scope of the appended claims.

Claims

1. An image processing method, characterized by, The method comprises the following steps: obtaining a vehicle use scene parameter corresponding to a to-be-processed surrounding environment image; determining a target feature acquisition model corresponding to the vehicle use scene of the to-be-processed surrounding environment image from each pre-trained alternative feature acquisition model according to the vehicle use scene parameter, wherein the vehicle use scene is high-speed driving, medium-speed driving, low-speed driving, driving on the outermost lane of the road, driving on the innermost lane of the road, or driving on the middle lane of the road, each pre-trained alternative feature acquisition model comprises an alternative feature acquisition model corresponding to each vehicle use scene, and the alternative feature acquisition model corresponding to the vehicle use scene is used for processing the surrounding environment image in the vehicle use scene to obtain an aerial view feature corresponding to the surrounding environment image in the vehicle use scene; inputting the to-be-processed surrounding environment image into the target feature acquisition model to obtain an aerial view feature corresponding to the to-be-processed surrounding environment image; wherein the input of the target feature acquisition model is the surrounding environment image, and the output is the aerial view feature corresponding to the surrounding environment image.

2. The image processing method of claim 1, wherein, The vehicle use scene parameter is a vehicle speed. The target feature acquisition model corresponding to the vehicle use scene of the to-be-processed surrounding environment image is determined from each pre-trained alternative feature acquisition model according to the vehicle use scene parameter, comprising: determining whether the vehicle speed corresponding to the to-be-processed surrounding environment image is greater than or equal to a first vehicle speed threshold value; if the vehicle speed corresponding to the to-be-processed surrounding environment image is greater than or equal to the first vehicle speed threshold value, it is determined that the vehicle use scene of the to-be-processed surrounding environment image is high-speed driving, and the alternative feature acquisition model corresponding to high-speed driving is determined as the target feature acquisition model corresponding to the vehicle use scene of the to-be-processed surrounding environment image.

3. The image processing method of claim 2, wherein, After determining whether the vehicle speed corresponding to the to-be-processed surrounding environment image is greater than or equal to the first vehicle speed threshold value, further comprising: if the vehicle speed corresponding to the to-be-processed surrounding environment image is less than the first vehicle speed threshold value, determining whether the vehicle speed corresponding to the to-be-processed surrounding environment image is greater than or equal to a second vehicle speed threshold value; if the vehicle speed corresponding to the to-be-processed surrounding environment image is greater than or equal to the second vehicle speed threshold value, it is determined that the vehicle use scene of the to-be-processed surrounding environment image is medium-speed driving, and the alternative feature acquisition model corresponding to medium-speed driving is determined as the target feature acquisition model corresponding to the vehicle use scene of the to-be-processed surrounding environment image.

4. The image processing method of claim 3, wherein, After determining whether the vehicle speed corresponding to the to-be-processed surrounding environment image is greater than or equal to the second vehicle speed threshold value, further comprising: if the vehicle speed corresponding to the to-be-processed surrounding environment image is less than the second vehicle speed threshold value, it is determined that the vehicle use scene of the to-be-processed surrounding environment image is low-speed driving, and the alternative feature acquisition model corresponding to low-speed driving is determined as the target feature acquisition model corresponding to the vehicle use scene of the to-be-processed surrounding environment image.

5. The image processing method of claim 4, wherein, Before obtaining the vehicle use scene parameter corresponding to the to-be-processed surrounding environment image, further comprising: performing the following operations for each vehicle use scene: obtaining a preset number of surrounding environment images corresponding to vehicle use scenarios and bird's eye view features corresponding to each of the surrounding environment images, wherein the vehicle use scenarios are high-speed driving, medium-speed driving or low-speed driving; training a machine learning model by taking each of the surrounding environment images and the bird's eye view features corresponding to each of the surrounding environment images as a training sample, to obtain an alternative feature acquisition model corresponding to the vehicle use scenarios.

6. The image processing method of claim 1, wherein, The vehicle use scenario parameter is a vehicle position. According to the vehicle use scenario parameter, a target feature acquisition model corresponding to the vehicle use scenario of the to-be-processed surrounding environment image is determined from the pre-trained alternative feature acquisition models, including: If the vehicle position corresponding to the to-be-processed surrounding environment image is the outermost lane, it is determined that the vehicle use scenario of the to-be-processed surrounding environment image is driving on the outermost lane of the road, and the alternative feature acquisition model corresponding to driving on the outermost lane of the road is determined as the target feature acquisition model corresponding to the vehicle use scenario of the to-be-processed surrounding environment image. If the vehicle position corresponding to the to-be-processed surrounding environment image is the innermost lane, it is determined that the vehicle use scenario of the to-be-processed surrounding environment image is driving on the innermost lane of the road, and the alternative feature acquisition model corresponding to driving on the innermost lane of the road is determined as the target feature acquisition model corresponding to the vehicle use scenario of the to-be-processed surrounding environment image. If the vehicle position corresponding to the to-be-processed surrounding environment image is the middle lane, it is determined that the vehicle use scenario of the to-be-processed surrounding environment image is driving on the middle lane of the road, and the alternative feature acquisition model corresponding to driving on the middle lane of the road is determined as the target feature acquisition model corresponding to the vehicle use scenario of the to-be-processed surrounding environment image.

7. The image processing method of claim 6, wherein, Before obtaining the vehicle use scenario parameter corresponding to the to-be-processed surrounding environment image, further comprising: For each vehicle use scenario, the following operations are performed: obtaining a preset number of surrounding environment images corresponding to vehicle use scenarios and bird's eye view features corresponding to each of the surrounding environment images, wherein the vehicle use scenarios are driving on the outermost lane of the road, driving on the innermost lane of the road or driving on the middle lane of the road; training a machine learning model by taking each of the surrounding environment images and the bird's eye view features corresponding to each of the surrounding environment images as a training sample, to obtain an alternative feature acquisition model corresponding to the vehicle use scenarios.

8. An image processing apparatus characterized by comprising: comprising: a parameter acquisition module configured to obtain a vehicle use scenario parameter corresponding to a to-be-processed surrounding environment image; The model determining module is configured to determine, according to the vehicle use scene parameter, a target feature acquisition model corresponding to a vehicle use scene of the to-be-processed surrounding environment image from each pre-trained candidate feature acquisition model; wherein the vehicle use scene is high-speed driving, medium-speed driving, low-speed driving, driving on an outermost lane of a road, driving on an innermost lane of a road, or driving on a middle lane of a road, and each pre-trained candidate feature acquisition model includes a candidate feature acquisition model corresponding to each vehicle use scene, and the candidate feature acquisition model corresponding to the vehicle use scene is used for processing a surrounding environment image in the vehicle use scene to obtain an overhead view feature corresponding to the surrounding environment image in the vehicle use scene. The feature acquisition module is configured to input the to-be-processed surrounding environment image into the target feature acquisition model to obtain an overhead view feature corresponding to the to-be-processed surrounding environment image. The input of the target feature acquisition model is a surrounding environment image, and the output is an overhead view feature corresponding to the surrounding environment image.

9. An electronic device, comprising: The electronic device includes: at least one processor; and a memory connected to the at least one processor in network communication; wherein the memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor to enable the at least one processor to execute the image processing method of any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer readable storage medium stores computer instructions for causing the processor to execute the image processing method of any one of claims 1-7 when executed.

Images