Rope skipping counting method, electronic equipment and storage medium

Abstract

The invention discloses a rope skipping counting method, electronic equipment and a storage medium. The rope skipping counting method comprises the following steps: acquiring attribute information of a plurality of target objects and configuration parameters of an image acquisition device; based on the attribute information and the configuration parameters of the plurality of target objects, determining a target position mode of the plurality of target objects; key point detection is carried out on a plurality of frames of target images, a key point set of a plurality of target objects is determined, and the plurality of frames of target images are acquired by the image acquisition device for each target object in the target station mode; and performing rope skipping statistics on the key point set of the plurality of target objects to obtain a rope skipping counting result of the plurality of target objects. According to the scheme, the accuracy of rope skipping counting results of the multiple target objects can be improved.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a rope skipping counting method, an electronic device and a storage medium. Background Art

[0002] The existing skipping rope counting scheme can only realize the skipping rope counting of a single target object. In addition, in the scenario where a large number of target objects need to be counted in time or at the same time, several counters are generally required to manually or semi-manually count and calculate the results in time. This type of skipping rope counting scheme has the disadvantages of easy counting errors, long statistical cycle, and unstable accuracy.

[0003] In view of the existing technical defects, how to provide a solution to improve the accuracy of rope skipping counting of multiple target objects is a technical problem that needs to be solved urgently by those skilled in the art. Summary of the invention

[0004] The present application at least provides a rope skipping counting method, an electronic device and a storage medium.

[0005] The present application provides a rope skipping counting method, comprising: acquiring attribute information of a number of target objects and configuration parameters of an image acquisition device; determining target standing modes of the number of target objects based on the attribute information and configuration parameters of the number of target objects; performing key point detection on a number of frames of target images to determine key point sets of the number of target objects, wherein the number of frames of target images are acquired by the image acquisition device from each target object in the target standing mode; performing rope skipping statistics on the key point sets of the number of target objects to obtain rope skipping counting results of the number of target objects.

[0006] The present application provides a rope skipping counting device, comprising: an acquisition module, a first determination module, a second determination module and a statistical module; the acquisition module is used to acquire attribute information of a number of target objects and configuration parameters of an image acquisition device; the first determination module is used to determine target standing modes of the number of target objects based on the attribute information and configuration parameters of the number of target objects; the second determination module is used to perform key point detection on a number of frames of target images to determine key point sets of the number of target objects, wherein the number of frames of target images are acquired by the image acquisition device from each target object in the target standing mode; the statistical module is used to perform rope skipping statistics on the key point sets of the number of target objects to obtain rope skipping counting results of the number of target objects.

[0007] The present application provides an electronic device, including a memory and a processor, wherein the processor is used to execute program instructions stored in the memory to implement the above-mentioned rope skipping counting method.

[0008] The present application provides a computer-readable storage medium on which program instructions are stored. When the program instructions are executed by a processor, the above-mentioned rope skipping counting method is implemented.

[0009] The above scheme dynamically determines the target position mode of several target objects based on the acquired attribute information and configuration parameters of several target objects. The target position mode can ensure that there is no occlusion between the target objects, and then key point detection is performed on several frames of target images collected from each target object in the target position mode. The key points of each target object in the key point set of the several target objects determined do not overlap, thereby improving the accuracy of the key point set of the several target objects determined, thereby performing rope skipping statistics on the key point set of the several target objects, and obtaining rope skipping counting results for the several target objects can improve the accuracy of the rope skipping counting results.

[0010] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] The drawings herein are incorporated into the specification and constitute a part of the specification. These drawings illustrate embodiments consistent with the present application and are used together with the specification to illustrate the technical solution of the present application.

[0012] Figure 1 This is a schematic diagram of the process of an embodiment of the rope skipping counting method of the present application. Figure 1 ; Figure 2 This is a schematic diagram of the process of an embodiment of the rope skipping counting method of the present application. Figure 2 ; Figure 3 This is a schematic diagram of the process of an embodiment of the rope skipping counting method of the present application. Figure 3 ; Figure 4 This is a schematic diagram of the target position in the first embodiment of the skipping rope counting method of the present application. Figure 1 ; Figure 5 This is a schematic diagram of the target position in the first embodiment of the skipping rope counting method of the present application. Figure 2 ; Figure 6 This is a structural schematic diagram of an embodiment of a skipping rope counting device of the present application; Figure 7 It is a structural schematic diagram of an embodiment of the electronic device of the present application; Figure 8 It is a structural diagram of an embodiment of a computer-readable storage medium of the present application. DETAILED DESCRIPTION

[0013] The scheme of the embodiment of the present application is described in detail below in conjunction with the drawings of the specification.

[0014] In the following description, for the purpose of explanation rather than limitation, specific details such as specific system structures, interfaces, and technologies are provided to facilitate a thorough understanding of the present application.

[0015] The term "and / or" in this article is only a description of the association relationship of associated objects, indicating that there may be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this article generally indicates that the objects associated before and after are in an "or" relationship. In addition, "many" in this article means two or more than two. In addition, the term "at least one" in this article means any combination of at least two of any one or more of a plurality of, for example, including at least one of A, B, and C, can mean including any one or more elements selected from the set consisting of A, B, and C.

[0016] The present application provides some rope skipping counting methods and rope skipping counting devices. The application scenarios of the rope skipping counting method include but are not limited to counting rope skipping for multiple target objects. The execution subject of the rope skipping counting method may be a rope skipping counting device. For example, the rope skipping counting device may be set in a terminal device or a server or other processing device, wherein the terminal device may be a device for rope skipping counting, a user equipment (UE), a mobile device, a user terminal, a terminal, a cellular phone, a cordless phone, a personal digital assistant (PDA), a handheld device, a computing device, a vehicle-mounted device, etc. In some possible implementations, the rope skipping counting method may be implemented by a processor calling a computer-readable instruction stored in a memory.

[0017] See also Figure 1 , Figure 1 This is a schematic diagram of the process of an embodiment of the rope skipping counting method of the present application. Figure 1 Specifically, the rope skipping counting method may include the following steps: Step S11: Acquire attribute information of several target objects and configuration parameters of the image acquisition device.

[0018] The target objects are at least one target object. The target objects are one or more. The target objects are animal bodies. The attribute information of the target objects represents the characteristics of the target objects. Specifically, the attribute information of the target objects represents the distribution statistics of the height information of each target object. Exemplarily, the attribute information of each target object includes the height information of the target object. The distribution statistics of the height information of each target object can be the mean, median, etc. corresponding to the height information of each target object. The present application takes the attribute information of the target objects as the mean corresponding to the height information of each target object as an example. The image acquisition device is used to acquire images of the target objects to obtain a plurality of frames of images, wherein each of the plurality of frames of images contains the area where each target object is located. The configuration parameters of the image acquisition device are used to represent at least one of the focal length, field of view angle, installation height, acquisition angle, imaging target surface size of the image sensor in the image acquisition device, minimum pixel size of the image sensor target surface in the image acquisition device, and resolution of the image sensor target surface in the image acquisition device when the image acquisition device acquires images of each target object. The installation height is the height of the image acquisition device from the ground. Among them, the configuration parameters of the image acquisition device are set according to the use requirements of the rope skipping counting, and the present application does not limit the specific values. Exemplarily, some configuration parameters of the image acquisition device of the present application are set as follows: the focal length is 2.3mm, the image sensor imaging target surface size is 9mm×5mm, the minimum pixel size of the sensor target surface is 2.9μm, the image sensor target surface resolution in the image acquisition device is 2688×1520, the field of view angle is 120°, the installation height is 4m, and the acquisition angle is 15° downward compared to the ground. It can be understood that under the configuration parameters of the image acquisition device, the image acquisition device can ensure that the images captured for several target objects can contain the areas where all target objects are located.

[0019] In some application scenarios, the method for obtaining the attribute information of several target objects in the above step S11 may be to obtain default height information, wherein the default height information is preset height information. In other application scenarios, the method for obtaining the attribute information of several target objects in the above step S11 may also be to first obtain the height information of each target object, and then perform statistical processing on the height information of each target object to obtain the attribute information of several target objects. In some application scenarios, the configuration parameters of the image acquisition device obtained in the above step S11 may be to directly obtain the preset parameters in the image acquisition device, and use the preset parameters as the configuration parameters of the image acquisition device.

[0020] Step S12: Determine target position modes of the target objects based on the attribute information and configuration parameters of the target objects.

[0021] The target position modes of the plurality of target objects include the rope skipping regions where the target objects are located in the image region corresponding to the field of view range of the image acquisition device. Specifically, the target position modes of the plurality of target objects are used to indicate the regions where the first preset parts of the target objects are located in the image region corresponding to the field of view range of the image acquisition device. The first preset parts may be the feet of the target objects.

[0022] The position constraint condition is set by the attribute information of several target objects and the configuration parameters of the image acquisition device. Based on the preset position condition, the rope skipping area where each target object is located is divided in the image area corresponding to the field of view angle range of the image acquisition device, and the rope skipping area where each target object is located is used as the position mode of several target objects. The position constraint condition is used to ensure that each target object in each frame image collected for several target objects in the target position mode is clear and there is no occlusion between the target objects. Among them, the position constraint condition includes the division condition of the positionable area in the image area corresponding to the field of view angle range and / or the division condition of the maximum object distance in the image area corresponding to the field of view angle range. Among them, when the image acquisition device acquires an image, some areas in the image may be distorted, and the positionable area is the area except the distorted area in the image area corresponding to the field of view angle range corresponding to the configuration parameters. The maximum object distance in the image area corresponding to the field of view angle range is used to indicate that the image collected for the target object at the maximum object distance is clear and the farthest distance between the target object and the vertical projection line of the image acquisition device.

[0023] It can be considered that by using attribute information and configuration parameters based on several target objects to determine the target positioning modes of several target objects, it can ensure that the target objects in each frame of target image collected for the target objects in the target positioning mode are clear and there is no occlusion between the target objects, thereby improving the accuracy of each frame of target image collected.

[0024] Step S13: perform key point detection on several frames of target images to determine key point sets of several target objects.

[0025] The target images of the plurality of frames are acquired by the image acquisition device from each target object in the target position mode. The key point sets of the plurality of target objects include a key point subset of each target object. The key point subset of the target object includes the target key points of the target object in each frame of the target image.

[0026] In some application scenarios, the above step S13 may be to first determine the segmentation frame of each target object based on the target position mode. The segmentation frame of each target object is used to segment several frames of target images to obtain a segmentation image set to which each target object belongs. The segmentation image set to which each target object belongs includes segmentation images obtained after segmentation of each frame of target image by the segmentation frame belonging to the target object. For each target object, key point detection is performed on each segmentation image to which the target object belongs to obtain a key point subset of the target object.

[0027] In other application scenarios, the key point detection of several frames of target images in the above step S13 may be to input the image features belonging to each frame of target image into the key point detection module for key point detection processing to obtain a subset of key points of each target object. A preset key point detection algorithm may be set on the key point detection module, wherein the preset key point detection algorithm may include but is not limited to an algorithm based on coordinate value regression, an algorithm based on instance segmentation, and an algorithm based on a heat map. Specifically, the algorithm based on coordinate value regression may be to directly regress the coordinate values ​​of the key points in the image features using a fully connected layer. Specifically, the algorithm based on instance segmentation may be that each instance object, that is, the target object, corresponds to its own key point, avoiding the pairing problem during multi-target key point detection. Detect instance objects in the target image, and generate a bounding box for each instance object, which may be a target box or the above segmentation box. Segment the image features corresponding to the area of ​​the target box and extract key points from it. Specifically, the algorithm based on the heat map can be to represent the coordinates of each key point in each target detection area in the target image with a probability map, and give a probability to each pixel position in the image corresponding to each target detection area, indicating the probability that the pixel point at the pixel position belongs to the key point of the corresponding category. Usually, the probability of the pixel point closer to the key point position is closer to 1, and the probability of the pixel point farther from the key point is closer to 0.

[0028] Step S14: rope skipping statistics are performed on the key point sets of several target objects to obtain rope skipping counting results of the several target objects.

[0029] The rope skipping statistics may be a statistical result obtained by counting the key point sets of several target objects according to a preset rope skipping rule. The rope skipping counting results of several target objects include the number of rope skipping belonging to each target object. The preset rope skipping rule is related to the key point changes of at least one preset part of the several target objects.

[0030] Specifically, the above step S14 can be that for each target object, a motion trajectory diagram of the target object is first drawn using the key point set of the target object. According to the preset motion change rule in the preset rope skipping rule, rope skipping statistics are performed on the motion trajectory diagram to obtain the number of rope skipping of the target object. Among them, the preset motion change rule is related to the change of the motion state of at least one preset part to which the target object belongs. For example, when the motion state of at least one preset part changes once, it is recorded as one rope skipping, and the number of rope skipping of the target object is increased by one.

[0031] The above scheme dynamically determines the target position mode of several target objects based on the acquired attribute information and configuration parameters of several target objects. The target position mode can ensure that there is no occlusion between the target objects, and then key point detection is performed on several frames of target images collected from each target object in the target position mode. The key points of each target object in the key point set of the several target objects determined do not overlap, thereby improving the accuracy of the key point set of the several target objects determined, thereby performing rope skipping statistics on the key point set of the several target objects, and obtaining rope skipping counting results for the several target objects can improve the accuracy of the rope skipping counting results.

[0032] See also Figure 2 , Figure 2 This is a schematic diagram of the process of an embodiment of the rope skipping counting method of the present application. Figure 2 .

[0033] In some embodiments, the configuration parameters include the focal length and the field of view of the image acquisition device, and the above step S12 may include the following steps: Step S21: Determine the target object distance based on the attribute information and the focal length. The target object distance is used to represent the distance of several target objects facing the image acquisition device and the farthest distance from the vertical projection line of the image acquisition device. Step S22: Determine the target position mode of several target objects based on the attribute information, the field of view and the target object distance.

[0034] The target object distance is the maximum distance between the target object and the vertical projection line of the image acquisition device at which the image captured of the target object at the maximum object distance is clear. The target object distance can be the above-mentioned maximum object distance. The vertical projection line of the image acquisition device is a projection line perpendicular to the ground.

[0035] The above step S21 may be to first use the focal length to determine the mapping relationship between the height information of the target object in the attribute information and the image height information of the target object in the image captured by the image acquisition device. The target height information is obtained, and the target height information is the image height information of the target object in the image captured by the image acquisition device that is clear and has the lowest height information. Based on the target height information, the image distance between the target object and the vertical projection line of the image acquisition device is determined. The image distance is mapped to the actual distance through the above mapping relationship, and the actual distance is used as the target object distance.

[0036] In some embodiments, the attribute information includes the physical size and minimum resolution corresponding to the target objects, the configuration parameters include the minimum pixel size of the image sensor in the image acquisition device, and the above step S21 may include the following steps: determining the product between the physical size and the focal length as the first product. Determining the product between the minimum pixel size and the minimum resolution as the second product. Determining the target object distance based on the ratio between the first product and the second product.

[0037] The actual size corresponding to the target object is the height information corresponding to several target objects. Specifically, the height information corresponding to several target objects can be the mean, median, etc. of the actual height of each target object. The minimum resolution corresponding to the target object is used to indicate the minimum number of pixels of the image captured by the image acquisition device occupied by any one of the several target objects. That is, it is required to cover as many target objects as possible within the field of view of the image acquisition device, and it is required that the number of pixels of the image captured by the image acquisition device occupied by each target object is greater than or equal to 200 pixels. In addition, in order to avoid mutual interference between adjacent target objects during rope skipping, the spacing value in the spacing information between adjacent target objects among several target objects is within the range of 1.3±0.1m. The minimum pixel size of the image sensor is the minimum pixel size of the target surface of the image sensor in the above-mentioned image acquisition device. For example, the minimum pixel size can be set to 2.9μm. The target object distance can be the above-mentioned maximum object distance.

[0038] The above step S21 can also be based on a preset proportional relationship to determine the initial object distance. The initial object distance is used to indicate the distance between the farthest target object within the field of view angle and the image acquisition device. Specifically, the preset proportional relationship can be that the ratio between the actual object size and the imaging size is equal to the ratio between the initial object distance and the focal length. Among them, the imaging size is the product between the minimum pixel size and the minimum resolution. The initial object distance is the ratio between the first product determined by the product between the actual object size and the focal length and the imaging size. The ratio between the first product and the imaging size is determined as the initial object distance. It can be understood that in this application, the ratio between the former and the latter is the numerator and the latter is the denominator, which will not be repeated later. Based on the vertical distance and the initial object distance, the target object distance is determined. Among them, the vertical distance and the target object distance are the right-angled sides of the right triangle, the initial object distance is the hypotenuse of the right triangle, and the vertical distance is the distance between the target object and the horizontal plane where the image acquisition device is located. It can be understood that the vertical distance, the target object distance and the initial object distance satisfy the Pythagorean theorem.

[0039] For example, the actual size is 1700mm, the imaging size is 200×0.0029, and the focal length is 2.3mm. At this time, the initial object distance determined based on the preset proportional relationship is 6.74m. The installation height is 4m, and the vertical distance is about 2.3m. At this time, based on the vertical distance and the initial object distance, the target object distance determined is about 6m. That is, under the specification conditions of the image acquisition device, if the facial pixel requirements are to be met, the maximum distance between the target object within the field of view angle and the vertical projection point of the optical center of the image acquisition device shall not exceed 6m.

[0040] In some application scenarios, the field of view angle of the image acquisition device is determined according to the image sensor imaging target surface size and focal length. Specifically, the lateral size of the image sensor imaging target surface size and the focal length satisfy a preset trigonometric function relationship. Specifically, the process of determining the field of view angle of the image acquisition device can refer to formula (1): Formula (1); Wherein, L may represent the lateral dimension of the image sensor imaging target surface size. For example, when the image sensor imaging target surface size is 9 mm×5 mm, the lateral dimension is 9 mm. f may represent the focal length of the image acquisition device. For example, the focal length may be 2.3 mm. θ may represent the field of view of the image acquisition device. For example, under the specific values ​​of the lateral dimension and the focal length of the above-mentioned image sensor imaging target surface size, the field of view angle obtained according to the above-mentioned formula (1) is 120°.

[0041] In other application scenarios, the spacing information between adjacent target objects among several target objects may be a preset spacing, or a spacing obtained based on historical spacing data statistics, or a spacing obtained through a preset algorithm based on the height and weight in the attribute information of several target objects. This application takes the spacing value in the spacing information between adjacent target objects among several target objects as an example within the range of 1.3±0.1m.

[0042] In other application scenarios, different categories of target objects occupy different numbers of pixels in the picture collected by the image acquisition device, all of which are preset values. Before determining the minimum resolution, the categories to which the target objects belong are first determined. In response to the categories to which the target objects belong being the target categories, the minimum resolution corresponding to the target categories is determined as the minimum resolution corresponding to the target objects. This application takes the example that the number of pixels of the picture collected by the image acquisition device occupied by each target object is greater than or equal to 200 pixels, that is, the minimum resolution is 200.

[0043] The above step S22 can be based on the field of view angle and the target object distance to determine the target positionable area within the field of view angle range. The target positionable area can be in the image area corresponding to the field of view angle range, and the radius is an approximate sector of the target object distance. It can be understood that the target positionable area and the distortion range constitute a sector within the field of view angle range. According to the attribute information, the target position of each target object is determined from the target positionable area.

[0044] It can be considered that the positionable area within the field of view angle range of the image acquisition device limited based on the target object distance can ensure that the clarity of the image acquisition device for capturing each target object meets the standard when the target position mode is determined by the positionable area, thereby improving the accuracy of key point detection based on subsequent images associated with the target position mode, and further improving the accuracy of rope skipping counting.

[0045] See also Figure 3 , Figure 3 This is a schematic diagram of the process of an embodiment of the rope skipping counting method of the present application. Figure 3 .

[0046] In some embodiments, the above step S22 may include the following steps: Step S31: Acquire several candidate arcs whose radius is less than or equal to the target object distance. Each candidate arc is an arc within the field of view angle and with the vertical projection point of the image acquisition device as the center. Step S32: In response to several candidate arcs satisfying preset conditions, select at least one target arc from the several candidate arcs. The preset conditions include that the first line and the second line do not intersect, the first line is the line between the top of the first target object and the optical center of the image acquisition device, the second line is the line between the bottom of the second target object and the optical center of the image acquisition device, and the first target object and the second target object are target objects in adjacent arcs. Step S33: Based on the arc length of each target arc, determine the target position mode of several target objects.

[0047] The top of the first target object can be the head of the first target object, or it can be the first preset area above the head of the first target object. The bottom of the second target object can be the feet of the second target object, or it can be the second preset area below the feet of the first target object. It can be understood that if you want to set as many standing arcs as possible in the target standing mode, the smaller the image area occupied by the first preset area and the second preset area. The first preset area and the second preset area are determined based on the rope skipping counting accuracy and are not limited here. Exemplarily, the top of the first target object can be the center point of the head of the first target object or a position away from the optical center. The bottom of the second target object can be the center point of the foot of the second target object or a position close to the optical center.

[0048] The above step S31 may be a method of obtaining candidate arcs by determining a number of candidate radii, each of which is less than or equal to the target object distance and satisfies the initial conditions. The initial conditions include that the difference between candidate radii of adjacent sizes is at least greater than or equal to the initial difference, and the initial difference is determined based on the difference between the arc lengths of two adjacent historical targets in the historical station data. The arcs in the image area within the field of view angle range and with a radius of each candidate radius are determined as each candidate arc.

[0049] See also Figure 4 , Figure 4 This is a schematic diagram of the target position in the first embodiment of the skipping rope counting method of the present application. Figure 1 .

[0050] like Figure 4As shown, point Z1 is the optical center of the image acquisition device. Z2 is the vertical projection point. a is the first connecting line. b is the second connecting line. c is the vertical projection line. d is the radius line corresponding to the target object distance. The first target object and the second target object are target objects on adjacent candidate arcs. Among them, the connecting line between the position of the first target object on the ground and the vertical projection point and the connecting line between the position of the second target object on the ground and the vertical projection point may not overlap. That is, the actual positions of the first target object and the second target object are on adjacent candidate arcs, but may not be on the extension line of the same radius. In the process of selecting at least one target arc from a number of candidate arcs, the actual position of the first target object and the actual position of the second target object are projected onto the extension line of the same radius to obtain the projection position of the first target object and the projection position of the second target object. The above-mentioned preset conditions also include that the first target object and the second target object are target objects on adjacent arcs, and the distance between the projection position of the first target object and the projection position of the second target object is greater than or equal to the target projection distance. The target projection distance is used to indicate that the first connecting line does not intersect the second connecting line and the minimum distance between the projection position of the first target object and the projection position of the second target object.

[0051] Specifically, the above step S32 may be to first determine the candidate radius with the same length as the target object distance as the target radius. The arc corresponding to the target radius is the last target arc. According to the above preset conditions, the target arcs adjacent to the last target arc are selected from the remaining candidate arcs in turn. The above preset conditions need to ensure that when the first target object and the second target object are in position, the first connecting line and the second connecting line do not form an intersection line, and the distance between the projection position of the first target object and the projection position of the second target object is greater than or equal to the target projection distance. The radius corresponding to each target arc length and the vertical projection point is each target radius.

[0052] It can be considered that at this time, it is determined that each target object in the target position mode is on the target arc, and there is no occlusion between the target objects, thereby improving the accuracy of the subsequent counting results of rope skipping based on the target position mode.

[0053] The above step S33 may be to determine the arc length of each target arc. Based on the spacing information between adjacent target objects among the target objects, the arc range of each target arc is divided to obtain the target position of each target object on each target arc.

[0054] Exemplarily, the target radius determined by the present application is respectively the last target radius of the same length as the target object distance and the first target radius determined based on the last target radius. Among them, the length of the last target radius is 6m, and the length of the first target radius is 3m. The arc lengths of the target arcs determined based on the above-mentioned field of view angle and target radius are approximately 6.28m and 12.56m, respectively. The spacing value in the spacing information between adjacent target objects among the above-mentioned several target objects is in the range of 1.3±0.1m as the preset spacing. Several target points are determined from each target arc at the preset spacing, and the spacing between adjacent target points is the above-mentioned preset spacing. For each target point, with the target point as the center of the circle, the circular area corresponding to the preset radius is used as the target position corresponding to the target point. The target position corresponding to each target point is the position of each target object.

[0055] It can be considered that determining the target position of each target object on each target arc using the spacing information between adjacent target objects and the arc length of each target arc can ensure that the target object in the image obtained by capturing the target object at the target position is clear and the target objects do not block each other.

[0056] In some embodiments, the attribute information includes the spacing information between adjacent target objects among the plurality of target objects, and the target position mode of the plurality of target objects includes the target position of each target object at each target arc. The above step S33 may include the following steps: obtaining the distortion range corresponding to each target arc. Determine the difference between each target arc and the arc corresponding to the distortion range corresponding to each target arc as the target range of each target arc. Each target range is a range that can be assigned as the position of the plurality of target objects. Determine each target position based on the target range and spacing information of each target arc.

[0057] In the image area corresponding to the field of view angle range, the larger the target radius corresponding to the target arc, the smaller the distortion range contained in the target arc. For each target arc, the distortion range corresponding to the target arc is subtracted from the target arc to obtain the target range of the target arc. Each target range is the range of possible positions in the target arc excluding the distortion range.

[0058] The distortion range corresponding to each target arc can be obtained by determining the distortion range corresponding to the target arc based on the field of view angle and the target radius. In the case of a fixed field of view angle, the distortion range corresponding to arcs of different radii is a preset range.

[0059] Based on the target range and spacing information of each target arc, the method of determining each target position can be to determine a number of target points from the target range of each target arc at the above preset spacing, and the spacing between adjacent target points is the above preset spacing. For each target point, with the target point as the center, the circular area corresponding to the preset radius is used as the target position corresponding to the target point. The target position corresponding to each target point is the position of each target object.

[0060] In other application scenarios, the preset spacing corresponding to each target radius is different. The spacing value in the spacing information between adjacent target objects among the above-mentioned several target objects is 1.3±0.1m, which is the initial preset spacing of the reference radius. Based on the size relationship between each target radius and the reference radius, the initial preset spacing of the reference radius is adjusted to obtain the target preset spacing of each target radius. With the target preset spacing corresponding to each target radius, several target points are determined from the target range of each target arc, and the spacing between adjacent target points is the target preset spacing of the same target radius. For each target point, with the target point as the center of the circle, the circular area corresponding to the preset radius is used as the target position corresponding to the target point. The target position corresponding to each target point is the position of each target object.

[0061] See also Figure 5 , Figure 5 This is a schematic diagram of the target position in the first embodiment of the skipping rope counting method of the present application. Figure 2 .

[0062] like Figure 5As shown, Z2 is the vertical projection point mentioned above. S1 is the first target arc. The target radius corresponding to the first target arc is r, and r is 3m. S2 is the last target arc. The target radius corresponding to the last target arc is R, and R is 6m. The target object distance is 6m. The field of view is 120°. In some application scenarios, for the first target arc, the distortion range corresponding to the target arc is the arc corresponding to the preset angles on both sides of the target arc, and the preset angle is 33°. The arc corresponding to the remaining 54° of the target arc is the target range of the target arc. In other application scenarios, for the last target arc, the distortion range corresponding to the target arc is the arc corresponding to the preset angles on both sides of the target arc, and the preset angle is 20°. The arc corresponding to the remaining 80° of the target arc is the target range of the target arc. It can be understood that the farther the target arc is from the vertical projection point, the smaller the preset angle. Among them, the spacing value in the spacing information between adjacent target objects among the above-mentioned several target objects is 1.3±0.1m, which is the initial preset spacing of the reference radius. Based on the size relationship between each target radius and the reference radius, the initial preset spacing of the reference radius is adjusted to obtain the target preset spacing of each target radius. It can be understood that because the area where the target object is located after diffusion with the target point of each target object as the center will also occupy the arc corresponding to part of the area or part of the angle in the target range of the target arc. Therefore, for the first target arc, the degree of distance between each target object in the target range of the target arc is less than or equal to 18°. For the last target arc, the degree of distance between each target object in the target range of the target arc is approximately 10°.

[0063] The preset target spacing in the first target arc S1 is 0.94m. The preset target spacing in the last target arc S2 is 1.05m. The arc lengths of the first target arc and the last target arc determined based on the above-mentioned field of view angle and target radius are approximately 6.28m and 12.56m respectively. Several target points are determined from the target range of each target arc with the target preset spacing corresponding to the above-mentioned target radius, and the spacing between adjacent target points is the target preset spacing of the same target radius. For each target point, with the target point as the center of the circle, the circular area corresponding to the preset radius is used as the target position corresponding to the target point. The target position corresponding to each target point is the position of each target object. Specifically, the target points in the first target arc S1 are P1, P2, P3, and P4 respectively. The target points in the last target arc S2 are P5, P6, P7, P8, P9, P10, P11, and P12 respectively.

[0064] It can be understood that the position of each target object is determined in the above manner, that is, the above target position is determined to be able to capture the target object at the target position so that the target object is clear in the image and the target objects do not block each other.

[0065] In some embodiments, the above step S13 may include the following steps: determining regions of interest of several frames of target images, each region of interest being a region where each target object is located in each frame of target image. Performing key point detection processing on each target object in the region of interest in each frame of target image to obtain an initial key point set of each target object. Coordinate mapping is performed on the initial key point set of each target object to obtain a key point set of each target object.

[0066] Several frames of target images are used to represent image data obtained by using an image acquisition device to acquire images of a target object in a target position. The initial key point set of each target object includes key point detection results of the target object in each frame of target image. The key point detection results in each frame of target image include key nodes of a preset part of the target object.

[0067] Each region of interest is the region where each target object is located in each frame target image. The above-mentioned determination of the region of interest of several frames of target images can be to segment each target object from each frame of target image to obtain the region where the target object is located in each frame of target image. For each target object, key point detection processing is performed on the target object in the region of interest in each frame of target image to obtain the initial key point set of the target object. Among them, the initial key point set of each target object is in its own coordinate system. Coordinate mapping of the initial key point set of each target object can ensure that the key point set of each target object is in the same coordinate system.

[0068] In some application scenarios, the target object is first detected by other feature recognition, and the other features may be features corresponding to preset parts of the target object. Then the key points of a single target object are analyzed and captured, and the key points of the target object are associated with the ID of the target object to obtain an initial key point set associated with each target object. Finally, the combined semantics of the key points in the initial key point set of each target object are analyzed and identified to obtain a motion trajectory diagram of the target object, thereby obtaining the skipping rope counting result of each target object based on the motion trajectory diagram. Exemplarily, the above step S13 may be to determine the key point set of each target object using a preset key detection model. The present application may adopt the regional multi-person pose estimation RMPE (AlphaPose) algorithm. Regional multi-person pose estimation can realize multi-person posture estimation and key point extraction. Specifically, a spatial exchange network (STN) is used to segment a high-quality pixel area corresponding to each target object from a certain pixel area of ​​each frame target image. Then the single person posture detection operator (SPPE) is used to extract the key points of the target object to obtain the initial key point set of the target object. Then, the initial key point set of each target object is mapped back to the original pixel coordinate system through the spatial exchange network (SDTN), and finally a key point set of each target object is formed.

[0069] Among them, for rope skipping, the present application marks the key nodes of preset parts of the target object under a preset number. Specifically, the preset parts include left wrist, right wrist, head, cervical spine, left elbow joint, right elbow joint, left knee, right knee, left foot, and right foot.

[0070] It can be considered that the present application first determines the region of interest of each target object, and then determines the initial key point set of each target object based on the region of interest, so as to achieve accurate positioning of the position of each target object to ensure the accuracy of the initial key point set determined for each target object, and then maps each initial key point set back to the original pixel coordinate system, so as to ensure the consistency and reference of the key point set of each target object, and facilitate subsequent comparison between the key point sets.

[0071] In other application scenarios, the preset key point detection model is a pre-trained preset key detection model. Obtain training data. Among them, obtaining training data includes: using a video acquisition front-end device to record a skipping video of multiple target objects, and decomposing the skipping video into multiple frames of continuous image streams. Mark the key nodes of the target object on the above continuous image stream, and the marked key nodes are used as training data for training the preset key detection model. The marked target is the skipping rope to which the target object belongs in the preset position of each target object and the top area of ​​the target object. Select a variety of key point detection networks, and use the prepared training data to continuously select the neural network to obtain a key point detection network with relatively accurate key detection results, and set the key point detection network as the preset key point detection model. Video acquisition and key node identification, using an image acquisition device to collect several frames of target images in the target position mode in real time, detect and track the target objects in each frame of the target image in real time, and assign an ID to each target object. For consecutive different target images, the ID of the same target object remains unchanged. In each frame of the target image, for all target objects, the preset key point detection model identifies the labeled targets of each target object to obtain several key point detection results, and the key point detection results of each target object are associated with the corresponding ID.

[0072] Specifically, the training step of the preset key point detection model includes: obtaining training data, wherein the training data includes a plurality of sample images and key point annotation results in each sample image. Inputting each sample image into the preset key point detection model to obtain the key point prediction results output by the preset key point detection model. Based on the difference between the key annotation results and the key point prediction results of each target object, adjusting the parameters of the preset key point detection model to obtain a trained preset key point detection model.

[0073] In some embodiments, the above step S14 may include the following steps: for each target object, the key point set of the target object is counted according to the target preset rule to obtain the rope skipping counting result of the target object. The target preset rule includes recording a rope skipping as one rope skipping when the key point of the target part in the key point set of the target object completes a motion state change, the target part includes the first preset part and / or the second preset part of the target object, the motion state change includes the first motion state change and / or the second motion state change, the first motion state change is the landing state, the off-the-ground state until the landing state, and the second motion state change is the second preset part meets the preset rope skipping rule.

[0074] For each target object, rope skipping statistics are performed on the key point set of the target object according to the target preset rules, and the rope skipping counting result of the target object can be obtained by first converting the key point set of each target object into a motion trajectory diagram corresponding to each target object. Based on the motion trajectory diagram corresponding to each target object, the motion state change of each target object is analyzed. The rope skipping counting result of each target object is determined by the relationship between the motion state change of each target object and the target preset rules.

[0075] Among them, the target preset rule includes recording one rope skipping when a key point of a target part in a set of key points of the target object completes a motion state change, the target part includes a first preset part and / or a second preset part of the target object, the motion state change includes a first motion state change and / or a second motion state change, the first motion state change is a landing state, a lift-off state until a landing state, and the second motion state change is that the second preset part satisfies the preset rope skipping rule.

[0076] In some application scenarios, when the target part in the target preset rule is the first preset part of the target object, the motion state change included in the target preset rule is the first motion state change, and the specific completion of the first motion state change is that the key point of the first preset part is in the landing state, the off-the-ground state until the landing state changes once. For example, the first preset part is the left sole and / or the right sole mentioned above. Exemplarily, the first motion state change completed in the trigger area of ​​the first preset part of the target object will be recorded as skipping rope once. Considering that when the target object is in a continuous skipping rope action, the first preset part may gradually deviate from the trigger area of ​​the target object during the jumping process, resulting in a misjudgment and non-counting phenomenon. The following algorithm logic is designed. When the coordinate values ​​of the center pixels of the first preset part, that is, the left and right sole areas, differ by ≥ the first preset pixel coordinate difference during the two consecutive landings, for example, the first preset pixel coordinate difference can be , when the user is currently in the landing state, the first preset part, that is, the pixel point set area of ​​the left and right soles of the feet, is used as the trigger area for the next judgment. The target image captured by the image acquisition device is composed of discrete pixels. Through a specific encoding method, a pixel point coordinate system is established for each frame of the target image. Each pixel point of the image can be identified as P (x, y), so the pixel point set M [P1, Pm] can identify any pixel area of ​​the target image. For the standing area of ​​each target object, the pixel point set of the circle area where the target object is located can be identified as the initial trigger area of ​​the human body, that is, when the target object is stationary, that is, the first preset part is in the landing state, the pixel point set area of ​​the left and right soles.

[0077] Exemplarily, for multiple consecutive frames of target images, the target object with the same ID is also the target object in the same domain of interest. The key point set of the adjacent frame target image is analyzed to obtain the motion trajectory diagram corresponding to the target object, and the motion trajectory diagram includes the key node motion trajectory of the target object. The ground in the target image is calibrated as the trigger area. When the distance between the first preset part and the trigger area is ≥10 pixels, it is judged to be off the ground. When the distance between the first preset part and the trigger area is ≤1 pixel, it is judged to be on the ground. The preset key points in two consecutive frames of target images are all within the quadrant above the trigger area, and are in a quasi-sinusoidal waveform motion within a certain set threshold pixel range. At the same time, when the left and right soles of the feet change from the ground state to the off-the-ground state and then to the ground state again, the skipping rope count is increased by one.

[0078] In other application scenarios, when the target part in the target preset rule is the second preset part of the target object, the motion state change included in the target preset rule is the second motion state change, and the key point of completing the second motion state change to the second preset part once satisfies the preset rope skipping rule. The preset rope skipping rule can be a preset action that can be achieved by the second preset part when completing one rope skipping. The preset action can be a circle action or a rotation action. For example, the second preset part is the above-mentioned left wrist and / or right wrist.

[0079] In other application scenarios, when the target part in the target preset rule is the first preset part of the target object and the second preset part of the target object, the motion state change included in the target preset rule is the first motion state change and the second motion state change. Then, the rope skipping counting results determined by the first motion state change and the second motion state change are averaged to obtain the final counting result, and the final counting result is used as the rope skipping counting result of the target object. It can be considered that the motion state change included in the target preset rule is the first motion state change and the second motion state change, which can avoid the situation where only jumping in place can be misjudged as rope skipping, thereby improving the accuracy of rope skipping counting.

[0080] It can be considered that the overall solution business of the present application can be an application system composed of a front-end device corresponding to the image acquisition device, a video storage device, an intelligent analysis module, and a performance statistics application system, which can improve the processing efficiency of the overall process of the rope skipping business. The present application can detect and track the continuous rope skipping actions of multiple target objects in real time, and perform key point analysis on multiple target objects at the same time. The key point information of each target object is associated and the corresponding motion trajectory diagram is drawn. Based on the motion trajectory diagram, irregular rope skipping actions such as false jumps and rope stepping can be effectively filtered out, thereby improving the accuracy of rope skipping counting.

[0081] In some embodiments, the key point set of each target object includes the key points of each target object and the skipping rope to which each target object belongs. The above step of performing skipping rope statistics on the key point set of the target object according to the target preset rule for each target object to obtain the skipping rope counting result of the target object may include the following steps: for each target object, performing skipping rope statistics on the key point set of the target object according to the target preset rule and the verification preset rule to obtain the skipping rope counting result of the target object. The verification preset rule includes that the first preset part of the target object is in a grounded state, and the coordinate difference between the key point of the skipping rope to which the target object belongs and the center point of the first preset part is greater than or equal to the preset value.

[0082] The key points of the skipping rope to which the target object belongs are the key points of the skipping rope of each target object. When skipping rope, the parabola vertex of the rope swinging process is identified as the key point of skipping rope. The specific preset verification rules include that when the left and right soles of the feet land, the difference between the pixel coordinate values ​​of the center point of the image area to which the first preset part belongs and the key point of the skipping rope is greater than or equal to the second preset pixel coordinate difference, wherein the second preset pixel coordinate difference can be 50.

[0083] Specifically, rope skipping statistics are performed on the key point set of the target object according to the preset target rules to obtain the rope skipping counting result of the target object, and the preset verification rules are used to determine whether each motion state change in the rope skipping counting result is valid. If the preset verification rules are met, it means that the rope skipping count is a valid count. If the preset verification rules are not met, it means that the rope skipping count is an invalid count.

[0084] It can be considered that the target preset rules and the verification preset rules are used to count the rope skipping of the target object in the target standing position, which can effectively filter out irregular rope skipping actions such as false jumps and rope stepping, thereby improving the accuracy of rope skipping counting.

[0085] In some embodiments, after the above step S14, the rope skipping counting method may further include the following steps: identifying each target object in a plurality of frames of target images to obtain a recognition result of each target object, and associating the recognition result of each target object with the rope skipping counting result of each target object to obtain an association result of each target object.

[0086] The recognition result of each target object is used to represent the associated information related to the target object. The associated information related to the target object may be the target object attribute information, specifically name, gender, school, etc. The associated result of each target object includes the skipping rope counting result corresponding to the specific target object.

[0087] Identify each target object in a plurality of frames of target images, and obtain the identification results of each target object, including: for each target object, first collect the identification image of the part to be identified of the target object. For example, the part to be identified of the target object is the head. In response to the successful similarity match between the identification image to which the target object belongs and the stored image in the database, the relevant information of the successfully matched stored image is used as the identification result of the target object. Associate the identification result of each target object with the skipping rope counting result of each target object to obtain the association result of each target object.

[0088] Exemplarily, the rope skipping count values ​​in the rope skipping count results of all target object IDs captured by the image acquisition device are obtained. Figure 5 As shown, the target object ID of the target object at point P1 can be 1. At the same time, for each target object ID, an image to be identified is collected and associated with the target object ID. Based on the image to be identified associated with each target object ID, it is compared with the base library to which the target object belongs to complete the identification of the target object, and finally the corresponding skipping rope count value is bound to the associated information of the target object identified in the recognition result, thereby completing the skipping rope performance data statistics. For example, the associated information of the target object can be the name Zhang San. The associated results of each target object are specifically the skipping rope count results of Zhang San.

[0089] It can be considered that the association result obtained by associating the recognition result of each target object with the rope skipping counting result in the present application can facilitate the statistics of the rope skipping counting result of the specific target object and improve the efficiency of rope skipping counting.

[0090] The above scheme dynamically determines the target position mode of several target objects based on the acquired attribute information and configuration parameters of several target objects. The target position mode can ensure that there is no occlusion between the target objects, and then key point detection is performed on several frames of target images collected from each target object in the target position mode. The key points of each target object in the key point set of the several target objects determined do not overlap, thereby improving the accuracy of the key point set of the several target objects determined, thereby performing rope skipping statistics on the key point set of the several target objects, and obtaining rope skipping counting results for the several target objects can improve the accuracy of the rope skipping counting results.

[0091] See also Figure 6 , Figure 6 It is a structural schematic diagram of an embodiment of the skipping rope counting device of the present application. The skipping rope counting device 60 includes an acquisition module 61, a first determination module 62, a second determination module 63 and a statistical module 64. The acquisition module 61 is used to acquire the attribute information of several target objects and the configuration parameters of the image acquisition device; the first determination module 62 is used to determine the target standing mode of several target objects based on the attribute information and configuration parameters of several target objects; the second determination module 63 is used to perform key point detection on several frames of target images to determine the key point sets of several target objects, and the several frames of target images are obtained by the image acquisition device for each target object in the target standing mode; the statistical module 64 is used to perform skipping rope statistics on the key point sets of several target objects to obtain the skipping rope counting results of several target objects.

[0092] The above scheme dynamically determines the target position mode of several target objects based on the acquired attribute information and configuration parameters of several target objects. The target position mode can ensure that there is no occlusion between the target objects, and then key point detection is performed on several frames of target images collected from each target object in the target position mode. The key points of each target object in the key point set of the several target objects determined do not overlap, thereby improving the accuracy of the key point set of the several target objects determined, thereby performing rope skipping statistics on the key point set of the several target objects, and obtaining rope skipping counting results for the several target objects can improve the accuracy of the rope skipping counting results.

[0093] Please refer to the skipping rope counting method for the functions performed by each module, which will not be repeated here.

[0094] See also Figure 7 , Figure 7: is a schematic diagram of the structure of an embodiment of an electronic device of the present application. The electronic device 70 includes a memory 71 and a processor 72, and the processor 72 is used to execute the program instructions stored in the memory 71 to implement the steps in the above-mentioned skipping counting method embodiment. In a specific implementation scenario, the electronic device 70 may include but is not limited to: a multi-camera device, a microcomputer, and a server. In addition, the electronic device 70 may also include a mobile device such as a laptop computer and a tablet computer, which is not limited here.

[0095] Specifically, the processor 72 is used to control itself and the memory 71 to implement the steps in the above-mentioned skipping rope counting method embodiment. The processor 72 can also be called a CPU (Central Processing Unit). The processor 72 may be an integrated circuit chip with signal processing capabilities. The processor 72 can also be a general-purpose processor, a digital signal processor (Digital Signal Processor, DSP), an application-specific integrated circuit (Application Specific Integrated Circuit, ASIC), a field-programmable gate array (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc. In addition, the processor 72 can be implemented by an integrated circuit chip.

[0096] The above scheme dynamically determines the target position mode of several target objects based on the acquired attribute information and configuration parameters of several target objects. The target position mode can ensure that there is no occlusion between the target objects, and then key point detection is performed on several frames of target images collected from each target object in the target position mode. The key points of each target object in the key point set of the several target objects determined do not overlap, thereby improving the accuracy of the key point set of the several target objects determined, thereby performing rope skipping statistics on the key point set of the several target objects, and obtaining rope skipping counting results for the several target objects can improve the accuracy of the rope skipping counting results.

[0097] See also Figure 8 , Figure 8 The computer-readable storage medium 80 stores program instructions 801, which implement the steps in any of the above-mentioned rope skipping counting method embodiments when executed by a processor.

[0098] The above scheme dynamically determines the target position mode of several target objects based on the acquired attribute information and configuration parameters of several target objects. The target position mode can ensure that there is no occlusion between the target objects, and then key point detection is performed on several frames of target images collected from each target object in the target position mode. The key points of each target object in the key point set of the several target objects determined do not overlap, thereby improving the accuracy of the key point set of the several target objects determined, thereby performing rope skipping statistics on the key point set of the several target objects, and obtaining rope skipping counting results for the several target objects can improve the accuracy of the rope skipping counting results.

[0099] In some embodiments, the functions or modules included in the device provided by the embodiments of the present disclosure can be used to execute the method described in the above method embodiments. The specific implementation can refer to the description of the above method embodiments, and for the sake of brevity, it will not be repeated here.

[0100] The above description of various embodiments tends to emphasize the differences between the various embodiments. The same or similar aspects can be referenced to each other, and for the sake of brevity, they will not be repeated herein.

[0101] In the several embodiments provided in the present application, it should be understood that the disclosed methods and devices can be implemented in other ways. For example, the device implementation described above is only schematic. For example, the division of modules or units is only a logical function division. There may be other division methods in actual implementation, such as units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, and the indirect coupling or communication connection of devices or units can be electrical, mechanical or other forms.

[0102] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0103] 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 application is essentially 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. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) or a processor to execute all or part of the steps of each implementation method of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program code.

Claims

1. A method for counting rope skipping, characterized in that: The method comprises: Acquire attribute information of several target objects and configuration parameters of an image acquisition device; Determining target position modes of the target objects based on the attribute information of the target objects and the configuration parameters; Performing key point detection on a plurality of frames of target images to determine a set of key points of the plurality of target objects, wherein the plurality of frames of target images are acquired by the image acquisition device from each of the target objects in the target station mode; Rope skipping statistics are performed on the key point sets of the plurality of target objects to obtain rope skipping counting results of the plurality of target objects.

2. The method according to claim 1, characterized in that The configuration parameters include the focal length and field of view of the image acquisition device, The determining the target position modes of the plurality of target objects based on the attribute information of the plurality of target objects and the configuration parameters includes: Determine a target object distance based on the attribute information and the focal length, wherein the target object distance is used to represent the farthest distance between the plurality of target objects toward the image acquisition device and a vertical projection line of the image acquisition device; Based on the attribute information, the field of view angle, and the target object distance, the target position modes of the plurality of target objects are determined.

3. The method according to claim 2, characterized in that The determining the target position modes of the plurality of target objects based on the attribute information, the field of view angle, and the target object distance includes: Acquire a plurality of candidate arcs whose radius is less than or equal to the target object distance, each of the candidate arcs being an arc within the field of view angle range and having a vertical projection point of the image acquisition device as a center; In response to the plurality of candidate arcs satisfying a preset condition, selecting at least one target arc from the plurality of candidate arcs, the preset condition comprising that a first connecting line and a second connecting line do not intersect, the first connecting line is a connecting line between a top of a first target object and an optical center of the image acquisition device, the second connecting line is a connecting line between a bottom of a second target object and an optical center of the image acquisition device, and the first target object and the second target object are target objects located in adjacent arcs; Based on the arc length of each of the target arcs, the target positions of the target objects are determined.

4. The method according to claim 3, characterized in that The attribute information includes the distance information between adjacent target objects among the plurality of target objects, and the target position modes of the plurality of target objects include the target position of each target object at each target arc. The determining the target positions of the plurality of target objects based on the arc lengths of the target arcs includes: Obtaining the distortion range corresponding to each of the target arcs; Determine the difference between each target arc and the arc corresponding to the distortion range corresponding to each target arc as the target range of each target arc, each target range being a range that can be allocated as the station position of the plurality of target objects; Based on the target range of each target arc and the spacing information, each target position is determined.

5. The method according to claim 2, characterized in that: The attribute information includes the physical size and minimum resolution corresponding to the plurality of target objects, the configuration parameters include the minimum pixel size of the image sensor in the image acquisition device, and the determining the target object distance based on the attribute information and the focal length includes: Determine the product of the physical size and the focal length as a first product; Determining a product of the minimum pixel size and the minimum resolution as a second product; The target object distance is determined based on a ratio between the first product and the second product.

6. The method according to any one of claims 1 to 5, characterized in that The rope skipping statistics are performed on the key point sets of the plurality of target objects to obtain rope skipping counting results of the plurality of target objects, including: For each of the target objects, rope skipping statistics are performed on the key point set of the target object according to the preset target rule to obtain the rope skipping counting result of the target object; Among them, the target preset rule includes recording one rope skipping when a key point of a target part in a set of key points of the target object completes a motion state change, the target part includes a first preset part and / or a second preset part of the target object, the motion state change includes a first motion state change and / or a second motion state change, the first motion state change is a landing state, a lift-off state to a landing state, and the second motion state change is that the second preset part satisfies the preset rope skipping rule.

7. The method according to claim 6, characterized in that The key point set of each target object includes key points of each target object and the skipping rope to which each target object belongs. For each target object, rope skipping statistics are performed on a key point set of the target object according to a preset target rule to obtain a rope skipping counting result of the target object, including: For each of the target objects, rope skipping statistics are performed on the key point set of the target object according to the target preset rules and the verification preset rules to obtain the rope skipping counting result of the target object. The verification preset rules include that a first preset part of the target object is in a grounded state, and the coordinate difference between the key point of the rope skipping to which the target object belongs and the center point of the first preset part is greater than or equal to a preset value.

8. The method according to any one of claims 1 to 5, characterized in that The step of performing key point detection on a plurality of frames of target images to determine a set of key points of the plurality of target objects includes: Determine regions of interest of the target images of the plurality of frames, each region of interest being a region where each target object is located in the target image of each frame; Performing key point detection processing on each target object in the region of interest in the target image of each frame to obtain an initial key point set of each target object; The initial key point sets of each target object are coordinate mapped to obtain the key point sets of each target object.

9. The method according to any one of claims 1 to 5, characterized in that: After performing rope skipping statistics on the key point sets of the plurality of target objects to obtain rope skipping counting results of the plurality of target objects, the method further includes: Identify each of the target objects in the plurality of frames of target images to obtain a recognition result of each of the target objects; The recognition result of each target object is associated with the rope skipping counting result of each target object to obtain the association result of each target object.

10. An electronic device, characterized in that: include: A memory and a processor, wherein the memory stores program instructions, and the processor retrieves the program instructions from the memory to execute the method according to any one of claims 1 to 9.

11. A computer-readable storage medium having program instructions stored thereon, characterized in that: When the program instructions are executed by a processor, they are used to implement the method according to any one of claims 1 to 9.

Images