WO2019047663A1

WO2019047663A1 - Video format-based end-to-end automatic driving data storage method and device

Info

Publication number: WO2019047663A1
Application number: PCT/CN2018/099391
Authority: WO
Inventors: 闫泳杉; 郁浩; 郑超; 唐坤; 张云飞; 姜雨
Original assignee: 百度在线网络技术（北京）有限公司
Priority date: 2017-09-05
Filing date: 2018-08-08
Publication date: 2019-03-14
Also published as: CN107767486A

Abstract

A video format-based end-to-end automatic driving data storage method and device, the method comprising: determining a video compression parameter and reading attitude data (S110); reading image data according to the order of timestamps of the attitude data (S120); and using the video compression parameter to store the video file in a pre-determined server according to the type of the attitude data outputted by the automatic driving system, and store the image data as a video file (S130). The method stores the read attitude data as image data according to the order of timestamps and storing same as a video file, and thus can reduce storage space occupied by the data, and also reduce the amount of access to the network I/O, to establish a better automatic driving data model, further improving the learning efficiency of deep learning in the field of automatic driving.

Description

Method and device for storing end-to-end automatic driving data based on video format

This patent application claims to be submitted on September 5, 2017, the application number is 201710792055.4, the applicant is Baidu Online Network Technology (Beijing) Co., Ltd., and the invention name is "a storage method based on video format for end-to-end automatic driving data. The priority of the Chinese Patent Application, the entire disclosure of which is incorporated herein by reference.

Technical field

The present invention relates to the field of computers, and in particular, to a method and apparatus for storing end-to-end automatic driving data based on a video format.

Background technique

With the rapid development of deep learning and the in-depth study of artificial intelligence, the automotive industry has undergone revolutionary changes. Automated driving through end-to-end deep learning is a major research direction in the field of automatic driving. In the prior art, an automatic driving system generally uses a model established by data acquired in real time in front, output steering angle, and speed to perform deep learning. The more data collected, the more favorable the generated model is for deep learning. However, these data usually need to be stored in a specific file, which requires a large storage space, which limits the development of deep learning in the field of automatic driving.

Summary of the invention

One of the technical problems solved by the present invention is that the data collected in front of the automatic driving system needs to occupy a large storage space.

According to an embodiment of an aspect of the present invention, a method for storing end-to-end automatic driving data based on a video format is provided, including:

Determining video compression parameters and reading gesture data;

Reading image data according to a time stamp sequence of the gesture data;

The image data is stored as a video file using the video compression parameters.

According to an embodiment of another aspect of the present invention, a storage device for end-to-end automatic driving data based on a video format is provided, including:

Means for determining video compression parameters and reading gesture data;

Means for reading image data in accordance with a time stamp of the gesture data;

Means for storing the image data as a video file using the video compression parameters.

Since the present embodiment stores the read posture data as image data in the order of the time stamp and stores it as a video file, the storage space occupied by the data can be reduced, and the amount of access of the network I/O can also be reduced to establish a more A good autonomous driving data model, which in turn improves the learning efficiency of deep learning in the field of automatic driving.

Those skilled in the art will appreciate that although the following detailed description is made with reference to the illustrated embodiments and drawings, the invention is not limited to these embodiments. Rather, the scope of the invention is intended to be limited the scope of the invention

DRAWINGS

Other features, objects, and advantages of the present invention will become more apparent from the Detailed Description of Description

1 shows a flow chart of a method for storing end-to-end autopilot data based on a video format in accordance with an embodiment of the present invention.

FIG. 2 is a flow chart showing a method for storing end-to-end automatic driving data based on a video format according to Embodiment 1 of the present invention.

FIG. 3 is a flow chart showing a method for storing end-to-end automatic driving data based on a video format according to Embodiment 2 of the present invention.

4 is a block diagram showing a storage device for end-to-end automatic driving data based on a video format in accordance with an embodiment of the present invention.

FIG. 5 is a block diagram showing a storage device for end-to-end automatic driving data based on a video format according to Embodiment 3 of the present invention.

Fig. 6 is a block diagram showing a storage device for end-to-end automatic driving data based on a video format proposed in Embodiment 4 of the present invention.

The same or similar reference numerals in the drawings denote the same or similar components.

Detailed ways

Before discussing the exemplary embodiments in more detail, it should be noted that some exemplary embodiments are described as a process or method depicted as a flowchart. Although the flowcharts describe various operations as a sequential process, many of the operations can be implemented in parallel, concurrently or concurrently. In addition, the order of operations can be rearranged. The process may be terminated when its operation is completed, but may also have additional steps not included in the figures. The processing may correspond to methods, functions, procedures, subroutines, subroutines, and the like.

By "computer device", also referred to as "computer" in the context, is meant an intelligent electronic device that can perform predetermined processing, such as numerical calculations and/or logical calculations, by running a predetermined program or instruction, which can include a processor and The memory is executed by the processor to execute a predetermined process pre-stored in the memory to execute a predetermined process, or is executed by hardware such as an ASIC, an FPGA, a DSP, or the like, or a combination of the two. Computer devices include, but are not limited to, servers, personal computers, notebook computers, tablets, smart phones, and the like.

The computer device includes a user device and a network device. The user equipment includes, but is not limited to, a computer, a smart phone, a PDA, etc.; the network device includes but is not limited to a single network server, a server group composed of multiple network servers, or a cloud computing based computer Or a cloud composed of a network server, wherein cloud computing is a type of distributed computing, a super virtual computer composed of a group of loosely coupled computers. Wherein, the computer device can be operated separately to implement the present invention, and can also access the network and implement the present invention by interacting with other computer devices in the network. The network in which the computer device is located includes, but is not limited to, the Internet, a wide area network, a metropolitan area network, a local area network, a VPN network, and the like.

It should be noted that the user equipment, the network equipment, the network, and the like are merely examples, and other existing or future possible computer equipment or networks, such as those applicable to the present invention, are also included in the scope of the present invention. It is included here by reference.

The methods discussed below, some of which are illustrated by flowcharts, can be implemented in hardware, software, firmware, middleware, microcode, hardware description languages, or any combination thereof. When implemented in software, firmware, middleware or microcode, the program code or code segments to carry out the necessary tasks can be stored in a machine or computer readable medium, such as a storage medium. The processor(s) can perform the necessary tasks.

The specific structural and functional details disclosed are merely representative and are for the purpose of describing exemplary embodiments of the invention. The present invention may, however, be embodied in many alternative forms and should not be construed as being limited only to the embodiments set forth herein.

It should be understood that although the terms "first," "second," etc. may be used herein to describe the various elements, these elements should not be limited by these terms. These terms are used only to distinguish one unit from another. For example, a first unit could be termed a second unit, and similarly a second unit could be termed a first unit, without departing from the scope of the exemplary embodiments. The term "and/or" used herein includes any and all combinations of one or more of the associated listed items.

It will be understood that when a unit is referred to as "connected" or "coupled" to another unit, it can be directly connected or coupled to the other unit, or an intermediate unit can be present. In contrast, when a unit is referred to as being "directly connected" or "directly coupled" to another unit, there is no intermediate unit. Other words used to describe the relationship between the units should be interpreted in a similar manner (eg "between" and "directly between" and "adjacent to" Than "directly adjacent to", etc.).

The terminology used herein is for the purpose of describing the particular embodiments, The singular forms "a", "an", It is also to be understood that the terms "comprising" and """ Other features, integers, steps, operations, units, components, and/or combinations thereof.

It should also be noted that in some alternative implementations, the functions/acts noted may occur in a different order than that illustrated in the drawings. For example, two figures shown in succession may in fact be executed substantially concurrently or sometimes in the reverse order, depending on the function/acts involved.

The invention is further described in detail below with reference to the accompanying drawings.

1 is a flow chart of a method of storing end-to-end autopilot data based on a video format, in accordance with one embodiment of the present invention.

As shown in FIG. 1, the video format-based end-to-end automatic driving data storage method according to this embodiment includes the following steps:

S110. Determine a video compression parameter and read the posture data.

S120. Read image data according to a timestamp sequence of the posture data.

S130. Store the image data as a video file using the video compression parameter.

The steps are further described in detail below.

In step S110, the video compression parameter is first determined. In this embodiment, the test data may be subjected to a compression test by a compression parameter, and then determined as a video compression parameter according to the compression ratio of the compression test. The compression parameters therein include, but are not limited to, at least one of a codec, an inter-frame allocation code rate (crf), or a color space.

After determining the video compression parameters, the gesture data is read. In the present embodiment, the attitude data output by the predetermined automatic driving system can be read in real time, and the posture data is stored as a time-stamped data sequence.

In step S120, after the posture data is read, the image data is sequentially read according to the time stamp of the posture data. In the present embodiment, the image data may be read in the order of the time stamp of the posture data, and the image data may be stored as an image data sequence in the order of the time stamp.

In step S130, the read image data can be stored as a video file using the video compression parameters described above. In this embodiment, the image data may be compression-stored into a video file by a predetermined video format, and each frame of the video file corresponds to an image in the image data.

Optionally, the embodiment further stores the video file on the predetermined server according to the type of the posture data output by the automatic driving system. The types of gesture data include, but are not limited to, speed data, steering angle data, road network data, and the like.

With the technical solution proposed in this embodiment, by storing the read posture data as image data in the order of time stamp and storing it as a video file, the storage space occupied by the data can be reduced, and the access of the network I/O can also be reduced. Volume to build a better autonomous driving data model, which in turn improves the learning efficiency of deep learning in the field of automatic driving.

Embodiment 1

In the prior art in the art, the image acquired by the sensor is stored in the HDF5 file for use by machine learning and control software. This method will result in an HDF5 file storing images that is too large and will significantly increase the overhead of network I/O, so the traditional data acquisition method is not conducive to deep learning of the automatic driving system.

Therefore, this embodiment proposes another storage method for the end-to-end automatic driving data based on the video format. As shown in FIG. 2, the method includes the following steps:

S210. Determine a video compression parameter.

When choosing compression parameters, consider the environment in which the autonomous driving system is located. For example, when driving on a sparsely populated highway, the color space can be used as a compression parameter due to less landscape changes on both sides of the road. For the color characteristics of snow, desert, forest, etc., the same in multiple image data. The color is uniformly compressed, and only the changes in the road surface are stored separately.

For example, when traveling in an urban area, an inter-frame allocation code rate can be used as a compression parameter. By dividing the code rate between frames, it is possible to analyze which are important frames and which are secondary frames, and important frames get more bytes. For example, an object that is not moving in the image or a moving object that is far away is set as a secondary frame, and only when the distance is less than the threshold, the moving object appears in the compression parameter as an important frame. This can give a clearer feeling and significantly reduce the size of the video file, because usually the human eye only pays attention to the moving object, and does not recognize the background.

S220. Read posture data.

The automatic driving system outputs a set of posture data every predetermined time, and the posture data usually includes image data, speed data, steering angle data, and road network data. This embodiment mainly reads image data therein.

S230. The image data is read in the order of time stamps of the posture data.

The attitude data output by the autopilot system is time stamped, which can be used to indicate the order in which the gesture data is generated, and the image data storage in a chronological order can more accurately characterize the image acquired by the autopilot system.

Therefore, in this embodiment, all the posture data are read in the order of time stamps to ensure that the posture data is consistent with the time stamp of the image data. After reading a predetermined number of image data, the image data is stored as a data sequence for subsequent steps to be called.

S240. Generate the video file into the video file in a predetermined format.

This embodiment generates a video file in the FFmpeg format. FFmpeg can be used to record, convert, digital audio, video, and convert these into streams. FFmpeg can not only compress multiple image data to generate video files, but also convert between multiple video formats. In addition, for selected videos, it is also possible to capture thumbnails of a specified time and acquire still and dynamic images.

The number of image data used to generate a video file is different each time according to different compression parameters. When the color space is used as the compression parameter, it is possible to compress 10,000 images each time to generate a 24 frame/second video file, the length of the video file is 7 minutes, the occupied space is generally 20-50M, and the original image is occupied by storage. The space is about 1G, and the compressed video file not only occupies less storage space, but also has low network I/O overhead.

In this embodiment, the attitude data output by the automatic driving system is compressed and stored as a video file according to a predetermined compression parameter and a video format, which can significantly reduce the storage space occupied by the posture data, and can also ensure the clarity of the stored video file. Integrity, therefore, can improve the depth learning efficiency of the automatic driving system.

Embodiment 2

In the prior art in the art, the image acquired by the sensor is stored in the HDF5 file for use by machine learning and control software. This method will cause the HDF5 file to store images to be too large, and will obviously increase the overhead of network I/O. Image storage will also result in too many files being stored, which is not conducive to editing and management, so the traditional data acquisition method is not conducive to automatic driving. Deep learning of the system.

Although the occupied storage space can be reduced by compressing the image, when these files need to be read, an additional decompression process is required, and it is difficult to improve the efficiency of deep learning. Therefore, this embodiment proposes a storage method for end-to-end automatic driving data based on a video format. As shown in FIG. 3, the method includes the following steps:

S310. Determine a video compression parameter and read the posture data.

Before determining the video compression parameters, different parameters can be used on the test data, such as codec, inter-frame allocation code rate, color space, etc., to compare the compression of these compression parameters and the image clarity after compression.

When choosing compression parameters, consider the environment in which the autonomous driving system is located. For example, when driving in an urban area, because there are many buildings, pedestrians, vehicles, etc. on both sides of the road, and the colors are different, the buildings are not moving, and some pedestrians and vehicles are not moving. Therefore, an inter-frame allocation code rate can be used as a compression parameter. By dividing the code rate between frames, it is possible to analyze which are important frames and which are secondary frames. A non-moving object in the image or a moving object farther away from the image is set as a secondary frame, and only when the distance is less than the threshold, the moving object appears in the compression parameter as an important frame. The image thus compressed can highlight a moving object, that is, an object that has an image for automatic driving, and other objects that do not move will not occupy more storage space. In contrast, the compression effects of the other two compression parameters are significantly worse, so for the road conditions in the urban area, the embodiment preferably uses the inter-frame allocation code rate as the compression parameter.

S320. Read image data according to timestamp order of the posture data.

The autopilot system outputs a set of pose data every predetermined time, and each set of pose data is time stamped, which can be used to indicate the order in which the pose data is generated, and the image data storage can be more accurately characterized in chronological order. The image captured by the autopilot system.

S330. Generate the video file by using the image data in a predetermined format.

In this embodiment, the image data is generated into a video file by using an AVC encoding format, and the video file has a code rate of 208 kbps, a frame rate of 14 fps, and a resolution of 448×336. When the inter-frame allocation code rate is used as the compression parameter, the length of the video file generated by compressing 10,000 images is 12 minutes, the occupied space is generally 40-70M, and the storage space occupied by the original image is about 1G, after compression. Video files not only take up less storage space, but also have lower network I/O overhead.

S340. Store the type of the gesture data output by the autopilot system of the video file on a predetermined server.

The type of the attitude data generally includes speed data, steering angle data, road network data, and the like. Therefore, in this embodiment, the attitude data is divided into two categories for storage, the first type is dynamic data, including speed data, steering angle data, motor vehicle data, etc., and the second is static data, including building data, real-time road condition data, Traffic signal data, etc. Video files stored in this category are easy to edit and manage, improving the efficiency of deep learning.

In this embodiment, the attitude data output by the automatic driving system is compressed and stored as a video file according to a predetermined compression parameter and a video format, which can significantly reduce the storage space occupied by the posture data, and can also ensure the clarity of the stored video file. Integrity, and easy to edit and manage, avoiding additional decompression process, thus improving the deep learning efficiency of the autopilot system.

4 is a block diagram of a storage device for end-to-end automatic driving data based on a video format, in accordance with one embodiment of the present invention.

As shown in FIG. 4, the video format-based end-to-end automatic driving data storage device (hereinafter referred to as "storage device") includes the following devices:

Means for determining video compression parameters and reading posture data (hereinafter referred to as "compressed reading device") 410;

Means for reading image data according to the time stamp of the posture data (hereinafter referred to as "image reading device") 420;

Means for storing the image data as a video file using the video compression parameter (hereinafter referred to as "video generating device") 430.

The device will be further described in detail below.

The video compression parameters are first determined by the compressed reading device 410. In the present embodiment, the test data may be subjected to a compression test by the compression reading device 410 through a compression parameter, and then determined as a video compression parameter according to the compression ratio of the compression test. The compression parameters therein include, but are not limited to, at least one of a codec, an inter-frame allocation code rate (crf), or a color space.

After determining the video compression parameters, the gesture data is read. In the present embodiment, the attitude data output by the predetermined automatic driving system can be read in real time by the compression reading device 410, and the posture data is stored as a time-stamped data sequence.

After the posture data is read, the image data is read by the image reading device 420 in accordance with the time stamp of the posture data. In the present embodiment, the image data may be read by the image reading device 420 in the order of the time stamp of the posture data, and the image data may be stored as an image data sequence in the order of the time stamp.

After the image data is read, the read image data can be stored as a video file by the video generating device 430. In the present embodiment, the image data may be compression-stored into a video file by the video generating device 430 through a predetermined video format, and each frame of the video file corresponds to one image in the image data.

Optionally, the embodiment further stores the video file on the predetermined server according to the type of the posture data output by the automatic driving system by the video generating device 430. The types of gesture data include, but are not limited to, speed data, steering angle data, road network data, and the like.

Embodiment 3

Therefore, this embodiment proposes another storage device for end-to-end automatic driving data based on the video format, as shown in FIG. 5, including the following devices:

Means for determining a video compression parameter (hereinafter referred to as "compression parameter determining device") 510;

a device for reading posture data (hereinafter referred to as "attitude data reading device") 520;

Means for reading image data in order of time stamps of the posture data (hereinafter referred to as "first image data reading device") 530;

Means for generating a video file by using the image data in a predetermined format (hereinafter referred to as "first video file generating device") 540;

When choosing compression parameters, consider the environment in which the autonomous driving system is located. For example, when driving on a sparsely populated highway, the color space can be used as a compression parameter due to less landscape changes on both sides of the road. For the color characteristics of snow, desert, forest, etc., the compression parameter determining device 510 will be more The same color in the image data is uniformly compressed, and only the changes in the road surface are separately stored.

For example, when traveling in an urban area, an inter-frame allocation code rate can be used as a compression parameter. By assigning a bit rate between frames, it is possible to analyze which are important frames and which are secondary frames, and important frames get more bytes. For example, an object that is not moving in the image or a moving object that is far away is set as a secondary frame, and only when the distance is less than the threshold, the moving object appears in the compression parameter as an important frame. This can give a clearer feeling and significantly reduce the size of the video file, because usually the human eye only pays attention to the moving object, and does not recognize the background.

The automatic driving system outputs a set of posture data every predetermined time, and the posture data usually includes image data, speed data, steering angle data, and road network data. The present embodiment mainly reads image data therein by the posture data reading device 520.

Therefore, in this embodiment, all the posture data are read by the first image data reading device 530 in the order of time stamps to ensure that the posture data coincides with the time stamp of the image data. After reading a predetermined number of image data, the image data is stored as a data sequence for subsequent steps to be called.

The number of image data used to generate a video file is different each time according to different compression parameters. When the color space is used as the compression parameter, the first video file generating device 540 can select to compress 10,000 images each time to generate a video file of 24 frames/second, and the length of the video file is 7 minutes, and the occupied space is generally 20 -50M, the original image occupies about 1G of storage space. The compressed video file not only occupies less storage space, but also has lower network I/O overhead.

Embodiment 4

Although the occupied storage space can be reduced by compressing the image, when these files need to be read, an additional decompression process is required, and it is difficult to improve the efficiency of deep learning. Therefore, the present embodiment proposes a storage device for end-to-end automatic driving data based on a video format, as shown in FIG. 6, including the following devices:

Means for determining video compression parameters and reading posture data (hereinafter referred to as "compression and reading device") 610;

Means for reading image data according to the time stamp of the posture data (hereinafter referred to as "second image data reading device") 620;

Means for generating a video file by using the image data in a predetermined format (hereinafter referred to as "second video file generating device") 630;

Means for storing the type of the posture data output by the automatic driving system of the video file on a predetermined server (hereinafter referred to as "classification storage device") 640.

Before determining the video compression parameters, the compression and reading device 610 can use different parameters on the test data, such as codec, inter-frame allocation code rate, color space, etc., to compare the compression and compression of these compression parameters. After the image clarity.

When choosing compression parameters, consider the environment in which the autonomous driving system is located. For example, when driving in an urban area, because there are many buildings, pedestrians, vehicles, etc. on both sides of the road, and the colors are different, the buildings are not moving, and some pedestrians and vehicles are not moving. Therefore, an inter-frame allocation code rate can be used as a compression parameter. By dividing the code rate between frames, it is possible to analyze which are important frames and which are secondary frames. A non-moving object in the image or a moving object farther away from the image is set as a secondary frame, and only when the distance is less than the threshold, the moving object appears in the compression parameter as an important frame. The thus compressed image can highlight a moving object, that is, an object that has an image for automatic driving, and other immovable objects do not occupy more storage space. In contrast, the compression effects of the other two compression parameters are significantly worse, so for the road conditions in the urban area, the embodiment preferably uses the inter-frame allocation code rate as the compression parameter.

Therefore, in this embodiment, all the posture data are read by the second image data reading device 620 in the order of time stamps to ensure that the posture data coincides with the time stamp of the image data. After reading a predetermined number of image data, the image data is stored as a data sequence for subsequent steps to be called.

In this embodiment, the second video file generating device 630 generates the video file by using the AVC encoding format, and the video file has a code rate of 208 kbps, a frame rate of 14 fps, and a resolution of 448×336. When the inter-frame allocation code rate is used as the compression parameter, the length of the video file generated by compressing 10,000 images is 12 minutes, the occupied space is generally 40-70M, and the storage space occupied by the original image is about 1G, after compression. Video files not only take up less storage space, but also have lower network I/O overhead.

The type of the attitude data generally includes speed data, steering angle data, road network data, and the like. Therefore, the embodiment stores the posture data into two categories by the classification storage device 640. The first category is dynamic data, including speed data, steering angle data, motor vehicle data, etc., and the second is static data, including building data. , real-time traffic data, traffic signal data, etc. Video files stored in this category are easy to edit and manage, improving the efficiency of deep learning.

It should be noted that the present invention can be implemented in software and/or a combination of software and hardware. For example, the various devices of the present invention can be implemented using an application specific integrated circuit (ASIC) or any other similar hardware device. In one embodiment, the software program of the present invention may be executed by a processor to implement the steps or functions described above. Likewise, the software program (including related data structures) of the present invention can be stored in a computer readable recording medium such as a RAM memory, a magnetic or optical drive or a floppy disk and the like. Additionally, some of the steps or functions of the present invention may be implemented in hardware, for example, as a circuit that cooperates with a processor to perform various steps or functions.

It is apparent to those skilled in the art that the present invention is not limited to the details of the above-described exemplary embodiments, and the present invention can be embodied in other specific forms without departing from the spirit or essential characteristics of the invention. Therefore, the present embodiments are to be considered as illustrative and not restrictive, and the scope of the invention is defined by the appended claims instead All changes in the meaning and scope of equivalent elements are included in the present invention. Any reference signs in the claims should not be construed as limiting the claim. In addition, it is to be understood that the word "comprising" does not exclude other elements or steps. A plurality of units or devices recited in the system claims can also be implemented by a unit or device by software or hardware. The first, second, etc. words are used to denote names and do not denote any particular order.

While the invention has been shown and described with reference to the embodiments of the embodiments of the invention The protection sought herein is set forth in the appended claims.

Claims

A storage method for end-to-end automatic driving data based on a video format, comprising:

Determining video compression parameters and reading gesture data;

Reading image data according to a time stamp sequence of the gesture data;

The image data is stored as a video file using the video compression parameters.
The method of claim 1 wherein the step of determining video compression parameters comprises:

The test data is subjected to a compression test by a compression parameter, and the video compression parameter is determined according to the compression ratio of the compression test.
The method of claim 2, the compression parameter comprising at least one of a codec, an inter-frame allocation code rate, or a color space.
The method of claim 1 wherein the step of reading the pose data comprises:

The attitude data output by the predetermined automatic driving system is read in real time, and the posture data is stored as a time-stamped data sequence.
The method according to claim 1, wherein the step of reading the image data according to the time stamp of the posture data comprises:

The image data is read in the order of the time stamp of the pose data, and the image data is stored as an image data sequence in the order of the time stamp.
The method of claim 1, the step of storing the image data as a video file using the video compression parameter further comprises:

The video file is stored on the predetermined server in accordance with the type of the gesture data output by the automatic driving system using the video compression parameter.
A storage device for end-to-end automatic driving data based on a video format, comprising:

Means for determining video compression parameters and reading gesture data;

Means for reading image data in accordance with a time stamp of the gesture data;

Means for storing the image data as a video file using the video compression parameters.
The apparatus according to claim 7, wherein said means for determining a video compression parameter and reading the attitude data comprises:

And a device for performing compression testing on the test data by a compression parameter, and determining the video compression parameter according to a compression ratio of the compression test.
The apparatus of claim 8, wherein in the means for determining a video compression parameter and reading the pose data, the compression parameter comprises at least one of a codec, an inter-frame allocation code rate, or a color space.
The apparatus according to claim 7, further comprising: in the means for determining a video compression parameter and reading the attitude data:

The means for reading the attitude data output by the predetermined automatic driving system in real time and storing the posture data as a time-stamped data sequence.
The apparatus according to claim 7, wherein the means for sequentially reading image data according to a time stamp of the gesture data comprises:

Reading image data in the order of time stamps of the posture data, and storing the image data as an image data sequence in the order of the time stamps
The apparatus according to claim 7, wherein said means for storing said image data as a video file using said video compression parameter comprises:

Means for storing the video file on a predetermined server in accordance with the type of gesture data output by the automatic driving system using the video compression parameter.
A computer readable storage medium storing computer code, the method of any one of claims 1 to 6 being executed when the computer code is executed.
A computer program product, when the computer program product is executed by a computer device, the method of any one of claims 1 to 6 being performed.
A computer device, the computer device comprising:

One or more processors;

a memory for storing one or more computer programs;

When the one or more computer programs are executed by the one or more processors, the one or more processors are caused to implement the method of any one of claims 1 to 6.