Target capture method and system in e-commerce live broadcast process, and related equipment

Abstract

The invention is suitable for the technical field of target capture, and provides a target capture method and system in an e-commerce live broadcast process and related equipment, and the method comprises the steps: obtaining a frame of image from a video stream of a live broadcast site shot by photographic equipment; determining a first tracking frame of the target anchor and a second tracking frame of the to-be-captured target based on a multi-target tracking algorithm, and calculating a position relationship between the first tracking frame and the second tracking frame in the target surrounding area; judging whether the position relation of the first tracking frame and the second tracking frame in the target surrounding area meets a preset condition or not; if yes, matching the image of the target commodity meeting the preset condition with an image in a preset image rule base, and outputting a matching result; and adjusting the shooting parameters and the shooting angle of the shooting equipment according to the matching result. The target capture commodity is calculated based on the multi-target tracking algorithm, the parameters of the photographic equipment are automatically adjusted, errors caused by manual adjustment are reduced, and the watching experience of audiences is improved.

Description

Technical Field

[0001] The present invention relates to the field of target capture technology, and in particular to a target capture method, system and related equipment in an e-commerce live broadcast process. Background Art

[0002] Visual information plays an important role in human perception of the surrounding environment. With the rapid development of artificial intelligence technology and deep learning technology, computer vision is gradually replacing human eyes. Computer vision technology mainly includes object detection, image classification, image segmentation, target tracking and video target segmentation, among which object detection, image classification and image segmentation are processed for static images. At present, the more common target tracking technology is able to capture the specific position and movement trajectory of the tracked target in continuous video frames, which is the basis for advanced visual tasks such as target posture understanding and behavior analysis. Target tracking is divided into single target tracking and multi-target tracking. Single target tracking predicts the position and scale information of the target in subsequent video frames based on the calibration information of a single target in the first frame of the video. Multi-target tracking aims to continuously predict the position, scale and ID of multiple targets in continuous video frames.

[0003] In the prior art, when using multi-target tracking technology to perform special shooting processing on a target, it is impossible to automatically lock a specific target. For example, in the process of e-commerce live streaming, when a user needs to display a certain product, the photographer needs to manually operate the camera to take a close-up shot of the target product and adjust the shooting angle and parameters according to the placement and size of the product. This is more dependent on the photographer's shooting skills and experience, and it is easy to have shooting errors due to the photographer's mistakes, thus affecting the audience's viewing experience. Summary of the invention

[0004] The embodiments of the present application provide a method, system and related equipment for capturing a target in an e-commerce live broadcast process, which are used to solve the problem of manually adjusting the photographic equipment to affect the audience's viewing experience.

[0005] A first aspect of an embodiment of the present application provides a method for capturing a target in an e-commerce live broadcast process, comprising:

[0006] Acquire a frame of image from a video stream of a live broadcast site captured by a photographic device, wherein the image includes a target anchor and at least one target to be captured, wherein the target to be captured includes a target commodity;

[0007] Determine a first tracking frame of the target anchor and a second tracking frame of the target to be captured based on a multi-target tracking algorithm, and calculate a positional relationship between the first tracking frame and the second tracking frame in a target enclosing area, where the target enclosing area includes the tracking frame of the target anchor and the tracking frames of all the targets to be captured;

[0008] Determining whether a positional relationship between the first tracking frame and the second tracking frame in the target enclosing area satisfies a preset condition;

[0009] If so, the image of the target product that meets the preset conditions is matched with the image in the preset image rule library, and the matching result is output;

[0010] The shooting parameters and shooting angle of the photographic device are adjusted according to the matching result.

[0011] Furthermore, the determining of the first tracking frame of the target anchor and the second tracking frame of the target to be captured based on the multi-target tracking algorithm includes:

[0012] The first tracking frame and the second tracking frame are rotation frames, and the rotation frames include four vertices and a center point.

[0013] Furthermore, the calculating the positional relationship between the first tracking frame and the second tracking frame in the target enclosing area includes:

[0014] Filter the vertices or center points of the first tracking frame, and determine the coordinates of the vertices or center points in the target enclosing area;

[0015] Filter the vertices or center points corresponding to the second tracking frame, and determine the coordinates of the vertices or center points in the target enclosing area;

[0016] The difference between the coordinates of the vertices or the center point of the first tracking frame in the target enclosing area and the coordinates of the vertices or the center point corresponding to the second tracking frame in the target enclosing area is calculated respectively.

[0017] Furthermore, the determining whether the positional relationship between the first tracking frame and the second tracking frame in the target enclosing area satisfies a preset condition includes:

[0018] The coordinates of the vertices or center points of the first tracking frame in the target enclosed area are compared with the coordinates of the vertices or center points of the second tracking frame in the target enclosed area. The product with the smallest comparison value meets the conditions.

[0019] Furthermore, the determining whether the positional relationship between the first tracking frame and the second tracking frame in the target enclosing area satisfies a preset condition includes:

[0020] If not, the four vertices and one center point of the second tracking frame are re-determined.

[0021] Furthermore, the image of the target product that meets the preset conditions is matched with the image in the preset image rule library, and the matching result is output, including:

[0022] An image rule library is constructed, and the similarity between the image of the target product that meets the preset conditions and the images in the image rule library is calculated to generate a matching result.

[0023] Furthermore, the building of the image rule library includes:

[0024] Obtain a panoramic image of the live broadcast site taken by a photographic device, the panoramic image including image information of all targets to be captured;

[0025] Generate image features and pixel parameters corresponding to the image information of all targets to be captured;

[0026] The corresponding shooting parameters and shooting angles are set according to the image information of all the targets to be captured, and an image rule library is generated based on the image features and pixel parameters as well as the corresponding shooting parameters and shooting angles.

[0027] Furthermore, the expression for calculating the similarity between the image of the target product that meets the preset conditions and the images in the image rule library is as follows:

[0028]

[0029] Where: x is the feature vector of the target product image, y i is the feature vector of the image in the image rule base.

[0030] A second aspect of the embodiment of the present application provides a target capture system for an e-commerce live broadcast process, including:

[0031] An image acquisition unit, configured to acquire a frame of image from a video stream of a live broadcast site captured by a photographic device, wherein the image includes a target anchor and at least one target to be captured, wherein the target to be captured includes a target commodity;

[0032] a position relationship calculation unit, configured to determine a first tracking frame of the target anchor and a second tracking frame of the target to be captured based on a multi-target tracking algorithm, and calculate a position relationship between the first tracking frame and the second tracking frame in a target enclosing area, wherein the target enclosing area includes the tracking frame of the target anchor and the tracking frames of all the targets to be captured;

[0033] A position relationship determination unit, configured to determine whether a position relationship between the first tracking frame and the second tracking frame in the target enclosing area satisfies a preset condition;

[0034] a matching result output unit, configured to match the image of the target product meeting the preset condition with an image in a preset image rule library, and output a matching result when the positional relationship between the first tracking frame and the second tracking frame in the target enclosing area meets the preset condition;

[0035] A shooting parameter and shooting angle adjustment unit is used to adjust the shooting parameters and shooting angle of the photographic device according to the matching result.

[0036] A third aspect of an embodiment of the present application provides a computer device, including a memory, a processor, and computer-readable instructions stored in the memory and executable on the processor, wherein when the processor executes the computer-readable instructions, the steps of the target capture method of the e-commerce live broadcast process as described in any one of the above items are implemented.

[0037] It can be seen from the above technical solutions that the embodiments of the present application have the following advantages:

[0038] The present application first obtains a frame of image from the video stream of the live broadcast site shot by the photographic device, the image includes the target anchor and at least one target to be captured, determines the first tracking frame of the target anchor and the second tracking frame of the target to be captured based on the multi-target tracking algorithm, and calculates the positional relationship between the first tracking frame and the second tracking frame in the target encirclement area; then determines whether the positional relationship between the first tracking frame and the second tracking frame in the target encirclement area meets the preset conditions, and if so, matches the image of the target product that meets the preset conditions to the image in the pre-set image rule library, and outputs the matching result; finally, adjusts the shooting parameters and shooting angle of the photographic device according to the matching result. The present application calculates the product to be captured based on the multi-target tracking algorithm, thereby automatically adjusting the parameters of the photographic device according to the image information of the product, reducing the errors of manual adjustment and improving the viewing experience of the audience. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 A schematic diagram of an embodiment of a method for capturing a target in an e-commerce live broadcast process in the present invention;

[0040] Figure 2 It is a flow chart of another embodiment of a method for capturing a target in an e-commerce live broadcast process in the present invention;

[0041] Figure 3 This is a structural block diagram of an embodiment of a target capture system for an e-commerce live broadcast process in the present invention;

[0042] Figure 4 The figure is a structural block diagram of a computer device in the present invention. DETAILED DESCRIPTION

[0043] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0044] In this embodiment, the target capture method in the e-commerce live broadcast process is used to automatically adjust the shooting angle and parameters of the camera to improve the viewing experience of the audience. The implementation method in this embodiment can be implemented in the system, on the server, or on the terminal, without specific limitation.

[0045] Embodiment 1

[0046] See also Figure 1 , an embodiment of a target capture method in an e-commerce live broadcast process of the present invention comprises the following steps:

[0047] S11. Acquire a frame of image from the video stream of the live broadcast site captured by the camera, the image includes the target anchor and at least one target to be captured, the target to be captured includes the target product;

[0048] In actual applications, when the e-commerce anchor is displaying and introducing the goods, the photographer will adjust the shooting parameters and shooting angles appropriately according to the situation to give the goods an appropriate close-up. In this embodiment, by calculating the distance between the anchor and the goods, and matching the pixels of the goods, the parameters and angles of the camera are automatically adjusted without the need for manual adjustment by the photographer. In addition, the photographic device is set to one, the live broadcast venue is fixed, and the goods to be displayed during the live broadcast are also fixed in a certain place. The camera (photographic device) can achieve uninterrupted shooting during the live broadcast, where any frame of the image in the video stream is randomly obtained, where the image includes the image of the anchor in the live broadcast and at least one target to be captured, and the target to be captured refers to all the goods to be displayed in the live broadcast venue.

[0049] S12. Determine a first tracking frame of the target anchor and a second tracking frame of the target to be captured based on a multi-target tracking algorithm, and calculate a positional relationship between the first tracking frame and the second tracking frame in a target enclosing area, where the target enclosing area includes the tracking frame of the target anchor and the tracking frames of all the targets to be captured;

[0050] In this embodiment, the multi-target tracking algorithm uses Faster R-CNN as the target detector and VGG16 as the feature extraction network, and identifies the target anchor and the target to be captured as two categories. In order to avoid the situation where the identified target is blocked, a rotating frame is set instead of a rectangle or a square as the tracking frame. The rotating frame includes four vertices and a center point. One target corresponds to a tracking frame. The size of the tracking frame changes with the size of the target. The tracking frame of the target anchor is set as the first tracking frame, and the tracking frame of the target to be captured is set as the second tracking frame.

[0051] Calculating the positional relationship between the first tracking frame and the second tracking frame in the target enclosing area includes the following steps:

[0052] 1. Filter the vertices or center points of the first tracking frame and determine the coordinates of the vertices or center points in the target enclosing area;

[0053] 2. Filter the vertex or center point corresponding to the second tracking frame, and determine the coordinates of the vertex or center point in the target enclosing area;

[0054] 3. Calculate the difference between the coordinates of the vertices or center points of the first tracking frame in the target enclosing area and the coordinates of the vertices or center points corresponding to the second tracking frame in the target enclosing area.

[0055] Specifically, the target encirclement area refers to the range covered by a frame of image after mask processing, including all products captured in the frame. Then, the vertex or center point coordinate displacement difference between the anchor's tracking frame and the tracking frame of all targets to be captured in the target encirclement area is calculated respectively. When using vertex calculation, one, two or four vertices in the same direction can be selected for calculation. The more vertices there are, the higher the accuracy of the calculation.

[0056] After calculating the displacement difference between the anchor's tracking frame and the tracking frame of a target to be captured, the displacement difference is added to the set distance dictionary. After calculating the displacement differences between the tracking frames of all targets to be captured and the anchor's tracking frame, the displacement differences are arranged in order of size, and the target to be captured with the smallest displacement difference with the anchor's tracking frame is determined as the target product. The smallest displacement difference means that the anchor is currently displaying the product, so the system controls the camera to adjust the shooting parameters and angles based on the image information of the product.

[0057] S13. Determine whether the positional relationship between the first tracking frame and the second tracking frame in the target enclosing area satisfies a preset condition;

[0058] S14. When the preset conditions are met, the image of the target product that meets the preset conditions is matched to the image in the preset image rule library, and the matching result is output;

[0059] To determine whether the preset conditions are met, the coordinates of the vertices or center points of the first tracking frame in the target enclosed area are compared with the coordinates of the vertices or center points of the second tracking frame in the target enclosed area, that is, the displacement difference. The product with the smallest ratio meets the conditions, that is, the target product. After determining the target product to be displayed by the anchor, the system immediately extracts a frame of the product, matches the frame with images in the preset image rule library, filters out images that meet the conditions with similarity to the frame, and then determines the shooting parameters and angles of the camera corresponding to the image that meets the conditions in the rule library, and finally feeds the corresponding parameters and angles back to the camera. For example, when the anchor wants to show a refrigerator, the image with the highest similarity to the refrigerator in the image rule library is matched, and then the image with the highest similarity is determined to correspond to the optimal shooting parameters and angles of the camera in the live broadcast venue.

[0060] When the preset condition is not met, the four vertices and one center point of the second tracking frame are re-determined, and the displacement difference is calculated based on the re-determined four vertices and the center point.

[0061] S15. Adjust the shooting parameters and shooting angle of the photographic device according to the matching result.

[0062] The photographic device automatically adjusts the camera position according to the target product matching result fed back by the system and the output optimal shooting parameters and shooting angle. For example, when the photographic device determines the target product to be captured, it obtains the frame image corresponding to the last captured target product, takes a certain pixel parameter as an example, calculates the pixel parameter based on the frame image, and compares the calculated pixel parameter with the pixel parameter matched to the image rule library. If the difference or ratio between the calculated pixel parameter and the pixel parameter in the image rule library is greater than 0, it means that the shape of the last captured target product is smaller than the shape of the current captured target product, and the camera is controlled to increase the distance from the product. If the difference or ratio between the calculated pixel parameter and the pixel parameter in the image rule library is less than 0, it means that the shape of the last captured target product is larger than the shape of the current captured target product, and the camera is controlled to reduce the distance from the product. If the difference or ratio between the calculated pixel parameter and the pixel parameter in the image rule library is approximately 0, it means that the shape of the last captured target product is not much different from the shape of the current captured target product, and the camera does not need to be adjusted. In this way, a very clear live broadcast picture can be displayed without the need for manual adjustment by the photographer, which effectively improves the viewing experience of the audience.

[0063] Embodiment 2

[0064] See also Figure 2 In the present invention, the image of the target product that meets the preset conditions is matched with the image in the preset image rule library, and the matching result is output, including the following steps:

[0065] S141. Obtain a panoramic image of the live broadcast site captured by a photographic device, the panoramic image including image information of all targets to be captured;

[0066] In this embodiment, during the live broadcast process, the shooting device performs a panoramic shot of the live broadcast site before real-time shooting, wherein the image obtained by the panoramic shooting includes the anchor and all the products to be displayed, and then the captured panoramic image is transmitted to the system, so that the system generates an image rule library.

[0067] S142. Generate image features and pixel parameters corresponding to the image information of all targets to be captured;

[0068] S143. Setting corresponding shooting parameters and shooting angles according to the image information of all targets to be captured, and generating an image rule library based on image features and pixel parameters and corresponding shooting parameters and shooting angles;

[0069] The computer is used to analyze the visual features of the image. Here, the deep learning method is used to obtain the image features corresponding to the image information of the target to be captured, including color, texture, shape and size, etc., and the feature set and feature sequence are established based on these features. Then, the corresponding pixel parameters are calculated for each target to be captured, where the pixel parameters include the number of pixels, pixel density, pixel size, bit depth, hue, saturation and brightness, and image data sets based on pixel parameters are established respectively.

[0070] Set corresponding shooting parameters and shooting angles according to the image information of all targets to be captured. For example, when the live broadcast venue and displayed goods are fixed, the camera is fixed at a certain place in the venue. At this time, the system determines the pixel parameters of the refrigerator. The camera can obtain the optimal parameters and angles for shooting the refrigerator in the venue based on the position of the refrigerator in the live broadcast venue and the relative position of the camera. Finally, a rule library for the refrigerator is generated based on the feature vector of the refrigerator and the corresponding shooting parameter angle. When the system tracks the refrigerator as the target product, it automatically adjusts the camera to match the corresponding data of the refrigerator.

[0071] S144. Calculate the similarity between the image of the target product that meets the preset conditions and the images in the image rule library to generate a matching result.

[0072] The expression for calculating similarity is as follows:

[0073]

[0074] Where: x is the feature vector of the image of the product that meets the preset conditions, y i is the feature vector of the image in the image rule base.

[0075] Specifically, when the system determines a target product that meets the similarity, it generates the best shooting parameters and angles based on the number of pixels, pixel density, pixel size, bit depth, hue, saturation and brightness of the product image. For example, when the system feeds back a matching result, the camera adjusts the camera parameters according to the number of pixels, pixel density and pixel size in the matching result, that is, the distance of the current camera from the current product is adjusted relative to the previous product, and then the camera's shooting angle is adjusted according to the bit depth, hue, saturation and brightness.

[0076] Based on the above-mentioned pre-set image rule library, the system reduces complex calculations. After determining the target product, the parameters that need to be adjusted for the camera can be obtained based on the rule library, effectively improving the efficiency of automatic adjustment of the camera.

[0077] Embodiment 3

[0078] See also Figure 3 , an embodiment of a target capture system for an e-commerce live broadcast process of the present invention comprises the following steps:

[0079] An image acquisition unit 101 is used to acquire a frame of image from a video stream of a live broadcast site captured by a photographic device, wherein the image includes a target anchor and at least one target to be captured, and the target to be captured includes a target commodity;

[0080] A position relationship calculation unit 102 is used to determine a first tracking frame of a target anchor and a second tracking frame of a target to be captured based on a multi-target tracking algorithm, and calculate a position relationship between the first tracking frame and the second tracking frame in a target enclosing area, where the target enclosing area includes the tracking frame of the target anchor and the tracking frames of all the targets to be captured;

[0081] A position relationship determination unit 103 is used to determine whether the position relationship between the first tracking frame and the second tracking frame in the target enclosing area meets a preset condition;

[0082] The matching result output unit 104 is used to match the image of the target product that meets the preset condition with the image in the preset image rule library when the positional relationship between the first tracking frame and the second tracking frame in the target enclosing area meets the preset condition, and output the matching result;

[0083] The shooting parameter and shooting angle adjustment unit 105 is used to adjust the shooting parameters and shooting angle of the photographic device according to the matching result.

[0084] For the specific definition of the target capture system, please refer to the definition of the target capture method above, which will not be repeated here. Each module in the above target capture system can be implemented in whole or in part by software, hardware and a combination thereof. The above modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory in the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0085] Embodiment 4

[0086] See also Figure 4 The present invention provides a computer device, including a memory, a processor, and computer-readable instructions stored in the memory and executable on the processor. When the processor executes the computer-readable instructions, the steps of the above method are implemented.

[0087] Those of ordinary skill in the art will appreciate that the units of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition of each example has been generally described in terms of function in the above description. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0088] In the embodiments provided by the present invention, it should be understood that the division of units is only a logical function division, and there may be other division methods in actual implementation, such as multiple units can be combined into one unit, one unit can be split into multiple units, or some features can be ignored. In addition, each functional unit in each embodiment of the present invention can be integrated into a processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units.

[0089] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-0nlyMemory), random access memory (RAM, RandomAccessMemory), mobile hard disk, magnetic disk or optical disk, etc., which can store program code.

[0090] It can be understood that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or replace some or all of the technical features therein by equivalents. These modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present invention, and they should all be included in the scope of the claims and specification of the present invention.

Claims

1. A target capture method for an e-commerce live broadcast process, characterized in that: include: Acquire a frame of image from a video stream of a live broadcast site captured by a photographic device, wherein the image includes a target anchor and at least one target to be captured, wherein the target to be captured includes a target commodity; Determine a first tracking frame of the target anchor and a second tracking frame of the target to be captured based on a multi-target tracking algorithm, and calculate a positional relationship between the first tracking frame and the second tracking frame in a target enclosing area, where the target enclosing area includes the tracking frame of the target anchor and the tracking frames of all the targets to be captured; Determining whether a positional relationship between the first tracking frame and the second tracking frame in the target enclosing area satisfies a preset condition; If so, the image of the target product that meets the preset conditions is matched with the image in the preset image rule library, and the matching result is output; The shooting parameters and shooting angle of the photographic device are adjusted according to the matching result.

2. The target capture method of the e-commerce live broadcast process according to claim 1 is characterized in that: The determining the first tracking frame of the target anchor and the second tracking frame of the target to be captured based on the multi-target tracking algorithm includes: The first tracking frame and the second tracking frame are rotation frames, and the rotation frames include four vertices and a center point.

3. The target capture method of the e-commerce live broadcast process according to claim 2 is characterized in that: The calculating the positional relationship between the first tracking frame and the second tracking frame in the target enclosing area includes: Filter the vertices or center points of the first tracking frame, and determine the coordinates of the vertices or center points in the target enclosing area; Filter the vertices or center points corresponding to the second tracking frame, and determine the coordinates of the vertices or center points in the target enclosing area; The difference between the coordinates of the vertices or the center point of the first tracking frame in the target enclosing area and the coordinates of the vertices or the center point corresponding to the second tracking frame in the target enclosing area is calculated respectively.

4. The target capture method of the e-commerce live broadcast process according to claim 3 is characterized in that: The determining whether the positional relationship between the first tracking frame and the second tracking frame in the target enclosing area meets a preset condition includes: The coordinates of the vertices or center points of the first tracking frame in the target enclosed area are compared with the coordinates of the vertices or center points of the second tracking frame in the target enclosed area. The product with the smallest comparison value meets the conditions.

5. The target capture method of the e-commerce live broadcast process according to claim 4 is characterized in that: The determining whether the positional relationship between the first tracking frame and the second tracking frame in the target enclosing area meets a preset condition includes: If not, the four vertices and one center point of the second tracking frame are re-determined.

6. The target capture method of the e-commerce live broadcast process according to claim 1 is characterized in that: The method of matching the image of the target product that meets the preset conditions with the image in the preset image rule library and outputting the matching result includes: An image rule library is constructed, and the similarity between the image of the target product that meets the preset conditions and the images in the image rule library is calculated to generate a matching result.

7. The target capture method of the e-commerce live broadcast process according to claim 6 is characterized in that: The step of constructing an image rule library comprises: Obtain a panoramic image of the live broadcast site taken by a photographic device, the panoramic image including image information of all targets to be captured; Generate image features and pixel parameters corresponding to the image information of all targets to be captured; The corresponding shooting parameters and shooting angles are set according to the image information of all the targets to be captured, and an image rule library is generated based on the image features and pixel parameters as well as the corresponding shooting parameters and shooting angles.

8. The target capture method of the e-commerce live broadcast process according to claim 6 is characterized in that: The expression for calculating the similarity between the image of the target product that meets the preset conditions and the images in the image rule library is as follows: Where: x is the feature vector of the target product image, y i is the feature vector of the image in the image rule base.

9. A target capture system for an e-commerce live broadcast process, characterized in that: include: An image acquisition unit, configured to acquire a frame of image from a video stream of a live broadcast site captured by a photographic device, wherein the image includes a target anchor and at least one target to be captured, wherein the target to be captured includes a target commodity; a position relationship calculation unit, configured to determine a first tracking frame of the target anchor and a second tracking frame of the target to be captured based on a multi-target tracking algorithm, and calculate a position relationship between the first tracking frame and the second tracking frame in a target enclosing area, wherein the target enclosing area includes the tracking frame of the target anchor and the tracking frames of all the targets to be captured; A position relationship determination unit, configured to determine whether a position relationship between the first tracking frame and the second tracking frame in the target enclosing area satisfies a preset condition; a matching result output unit, configured to match the image of the target product meeting the preset condition with an image in a preset image rule library, and output a matching result when the positional relationship between the first tracking frame and the second tracking frame in the target enclosing area meets the preset condition; A shooting parameter and shooting angle adjustment unit is used to adjust the shooting parameters and shooting angle of the photographic device according to the matching result.

10. A computer device comprising a memory, a processor, and computer-readable instructions stored in the memory and executable on the processor, characterized in that: When the processor executes the computer-readable instructions, it implements the steps of the target capture method of the e-commerce live broadcast process as described in any one of claims 1 to 8.

Images