CN116229582B - Logistics unmanned aerial vehicle and man-machine logistics interactive system based on human body gesture recognition - Google Patents

Abstract

The invention belongs to the technical field of logistics drones, and in particular relates to a logistics drone based on human gesture recognition and a human-machine logistics interaction system. The invention can recognize the human body posture of the cargo owner, judge whether the cargo owner is in the posture of receiving goods, and then determine the way of unloading. Combined with the setting of unloading parameters, the drone can be safely released after hovering or carrying the goods to the ground. The restriction on the goods can not only ensure that the goods are in good condition, but also ensure that the owner can gradually adapt to the weight of the goods when receiving the goods, and avoid the situation where the goods hurt the owner. During this process, the owner's action information is used to predict its posture. Realize the purpose of independently determining the way of unloading goods, and at the same time adjust the hovering position of the drone in real time according to the predicted attitude, so that the owner of the goods can take away the goods more conveniently.

Description

A logistics drone and human-machine logistics interaction system based on human body gesture recognition

technical field

The invention belongs to the technical field of logistics drones, and in particular relates to a logistics drone based on human gesture recognition and a human-machine logistics interaction system.

Background technique

With the continuous development of logistics technology, when people distribute goods, they are not limited to ground distribution. For some small goods, drones can be used for distribution operations. Compared with ground distribution, the distribution process The advantage is that it can ignore most of the terrain, and its delivery speed is also higher than that of the ground delivery, and in the process of flying the drone, of course, in order to ensure that the goods can be safely delivered to the owner, the corresponding face recognition system must be To be equipped, in order to avoid the phenomenon of misdelivery of goods.

When the existing logistics drones deliver goods, they usually place the goods on the ground after checking the face information. The identification method, after the face recognition verification is passed, the drone receives the confirmation command, lands on the ground near the geographical coordinates corresponding to the recipient, releases the delivered goods, completes the delivery task and takes off for the return journey, not only the delivery method Single, and the ground environment of the delivery point is an uncontrollable area, which is easy to cause damage to some fragile goods. At this time, it is more suitable for the owner to pick up the goods directly from the drone. Secondly, the conventional drone , it needs to be shut down, and after the unloading is completed, it needs to be started again. The take-off stage consumes a lot of power of the UAV, and the battery's own battery life has a certain impact. The logistics drone and its interactive system that choose the delivery method.

Contents of the invention

The purpose of the present invention is to provide a logistics UAV based on human body posture recognition and a human-machine logistics interaction system, which can recognize the human body posture of the cargo owner, determine the hovering position of the drone, and can also select the ground landing position according to the human body posture of the cargo owner. It improves the flexibility of the delivery method, and no matter what the delivery method is, the UAV does not need to be shut down, and the impact on the battery is small.

The technical scheme that the present invention takes is specifically as follows:

A human-computer logistics interaction system based on human gesture recognition, including a data acquisition module, a feature extraction module, a primary judgment module, an unloading judgment module, a motion capture module, a feature analysis module, a prediction module, a secondary judgment module, and a central control module ;

The data collection module is used to collect cargo information and the face image of the owner, and mark the face image of the owner as a standard image;

The feature extraction module is used to extract the human body image of the cargo owner, wherein the human body image includes a face image and a joint image;

The first-level judgment module is used to identify the human body image of the cargo owner, and determine the unloading method according to the recognition result, wherein the unloading method includes hovering in the air to pick up the goods and ground unloading;

The unloading judgment module is used to set the unloading parameters according to the unloading mode, and when the unloading parameters are satisfied, the restriction on the goods is released;

The motion capture module is used to collect the motion information of the owner, decode the motion information, output multiple frames of continuous images, and mark them as images to be analyzed;

The feature analysis module is used to obtain all the images to be analyzed, and input the images to be analyzed into the trend analysis model, and output the movement trend of the owner;

The prediction module is used to predict the arm posture of the cargo owner according to the movement trend, and output it as the predicted posture;

The secondary judgment module is used to obtain the predicted attitude, and adjust the hovering position of the drone according to the predicted attitude;

The central control module is used to send and receive transfer information among the data collection module, feature extraction module, primary judgment module, motion capture module, feature analysis module, prediction module and secondary judgment module.

In a preferred solution, it also includes a communication module, which is used to remind the owner of the cargo to choose the delivery method;

If the cargo owner chooses to hover to pick up the cargo, the drone will hover in front of the cargo owner;

If the cargo owner chooses to drop the cargo on the ground, the drone will place the cargo on the ground;

If the owner of the goods does not respond, the body image of the owner is recognized by the first-level judgment module, and the delivery method is automatically determined according to the recognition result.

In a preferred solution, the feature extraction module includes a first execution unit and a second execution unit, and the execution priority of the first execution unit is higher than that of the second execution unit;

When the first execution unit executes, it captures the facial image of the cargo owner and uploads the facial image to the primary judgment module;

When the second execution unit executes, it captures the joint image of the consignor and uploads the joint image to the primary determination module, wherein the joint image includes shoulder joint features, elbow joint features and wrist joint features.

In a preferred solution, the primary judgment module includes a first judgment unit and a second judgment unit, and the execution priority of the first judgment unit is higher than that of the second judgment unit;

The first judging unit is used to judge whether the facial image is consistent with a standard image;

If they are consistent, it is determined that the information of the cargo owner is accurate, and the drone is hovered directly in front of the cargo owner;

If inconsistent, it is determined that the owner information is inconsistent, and the second determination unit is not executed;

The second judging unit is used to judge the positional relationship of the shoulder joint features, elbow joint features and wrist joint features;

If the wrist joint feature is located above the elbow joint feature and the elbow joint feature is located above the shoulder joint feature, it is determined that the cargo owner chooses to hover in the air to pick up the goods;

If the shoulder joint feature is located below the elbow joint feature, it is determined that the cargo owner chooses to drop the cargo on the ground.

In a preferred solution, the unloading parameter is set as a pressure change parameter between the cargo and the UAV, and the pressure change parameter is set according to the weight of the cargo, wherein the value of the pressure change parameter is greater than half of the weight of the cargo one.

In a preferred solution, a standard posture image is preset in the secondary judgment module, and is compared with it after obtaining the predicted posture;

If the predicted attitude is consistent with the standard attitude image, the UAV hovers at the designated position;

If the predicted attitude is inconsistent with the standard attitude image, a deviation distance is generated, and the hovering position of the drone is adjusted to the position of the predicted attitude.

In a preferred solution, a voice broadcast module is also included, and the voice broadcast module is used to play reminder acceptance sentences.

The present invention also provides a logistics drone based on human body posture recognition, which is applied to the above-mentioned human-machine logistics interaction system based on human body posture recognition, including:

A fuselage, a clamping mechanism for carrying cargo is provided under the fuselage, a tension sensor is provided on the clamping mechanism, and the tension sensor is used to detect the tension value between the clamping mechanism and the loaded cargo;

Wherein, after the first-level determination module determines the unloading mode, the unloading determination module sets the unloading parameter according to the unloading mode, and the unloading parameter is the pressure change value between the clamping mechanism and the loaded goods. When the tension sensor detects that the force of the clamping mechanism satisfies the delivery parameters, the clamping mechanism releases the clamping of the cargo.

In a preferred solution, the clamping mechanism includes a clamping plate and a two-way telescopic rod, two clamping plates are provided, both of the two clamping plates are arranged under the fuselage, and the two-way telescopic The rod is fixedly installed at the bottom end of the fuselage, and the two ends of the two-way telescopic rod are respectively fixedly connected with the top ends of the two clamping plates.

In a preferred solution, the clamping plates are all arranged in an L shape, and an outer cargo-carrying box is arranged between the two clamping plates, and a clamping box is provided on both sides of the cargo-carrying outer box. The corresponding snap-in notch of the holding plate.

The technical effect that the present invention obtains is:

The invention can recognize the human body posture of the cargo owner, judge whether the cargo owner is in the posture of receiving goods, and then determine the way of unloading. Combined with the setting of unloading parameters, the drone can be safely released after hovering or carrying the goods to the ground. The restriction on the goods can not only ensure that the goods are in good condition, but also ensure that the owner of the goods can gradually adapt to the weight of the goods when receiving the goods, and avoid the situation where the goods hurt the owner. Prediction, realizing the purpose of independently determining the delivery method, and at the same time adjusting the hovering position of the drone in real time according to the predicted attitude, so that the owner can take the goods more conveniently, and no matter what the delivery method is, the drone will not Need to shut down, less impact on the battery.

Description of drawings

Fig. 1 is a system block diagram provided by the present invention;

Fig. 2 is a system operation diagram provided by the present invention;

Fig. 3 is a schematic diagram of the overall structure of the logistics drone provided by the present invention.

In the accompanying drawings, the list of parts represented by each label is as follows:

body;

Clamping mechanism; 201, clamping plate; 202, two-way telescopic rod.

Detailed ways

In order to make the above objects, features and advantages of the present invention more obvious and comprehensible, specific implementations of the present invention will be described in detail below in conjunction with the accompanying drawings.

In the following description, a lot of specific details are set forth in order to fully understand the present invention, but the present invention can also be implemented in other ways different from those described here, and those skilled in the art can do it without departing from the meaning of the present invention. By analogy, the present invention is therefore not limited to the specific examples disclosed below.

Second, "one embodiment" or "an embodiment" referred to herein refers to a specific feature, structure or characteristic that may be included in at least one implementation of the present invention. "In a preferred embodiment" appearing in different places in this specification does not all refer to the same embodiment, nor is it a separate or selective embodiment that is mutually exclusive with other embodiments.

Secondly, the present invention is described in detail in conjunction with schematic diagrams. When describing the embodiments of the present invention in detail, for the convenience of explanation, the cross-sectional view showing the structure of the device will not be partially enlarged according to the general scale, and the schematic diagram is only an example, and it should not be used here. Limit the scope of protection of the present invention. In addition, the three-dimensional space dimensions of length, width and depth should be included in actual production.

Please refer to Fig. 1 and Fig. 2, the present invention provides a human-computer logistics interaction system based on human gesture recognition, including a data collection module, a feature extraction module, a primary judgment module, an unloading judgment module, a motion capture module, a feature Analysis module, prediction module, secondary judgment module and central control module;

The data acquisition module is used to collect cargo information and the face image of the owner, and calibrate the face image of the owner as a standard image;

The feature extraction module is used to extract the human body image of the cargo owner, wherein the human body image includes a face image and a joint image;

The first-level judgment module is used to identify the human body image of the cargo owner, and determine the delivery method according to the recognition result, wherein the delivery method includes hovering in the air to pick up the goods and ground delivery;

The unloading judgment module is used to set the unloading parameters according to the unloading method, and release the restriction on the goods when the unloading parameters are met;

The motion capture module is used to collect the motion information of the cargo owner, decode and process the motion information, output multiple frames of continuous images, and calibrate them as images to be analyzed;

The feature analysis module is used to obtain all the images to be analyzed, input the images to be analyzed into the trend analysis model, and output the movement trend of the owner;

The prediction module is used to predict the arm posture of the cargo owner according to the movement trend, and output it as the predicted posture;

The secondary judgment module is used to obtain the predicted attitude, and adjust the hovering position of the drone according to the predicted attitude;

The central control module is used to send and receive the transfer information between the data acquisition module, feature extraction module, first-level judgment module, motion capture module, feature analysis module, prediction module and second-level judgment module.

As mentioned above, in the logistics system, the traditional delivery method is mostly vehicle delivery, but the urban roads are complex, which will lead to increased delays in delivery of goods, but with the rapid development of drone technology, it has been able to deliver to some small The delivery of goods can increase the efficiency of goods delivery accordingly, and can also reduce the corresponding labor costs. In the process of delivering goods, UAVs often need to place the goods on the ground, and also need to stop their operation to avoid unmanned If the propeller of the drone hurts the consignee or consignor, this will undoubtedly increase the frequency of opening and closing of the drone, which will increase the energy consumption of the drone and reduce the endurance of the drone. Based on this, this The scheme adopts the method of human body posture recognition for the cargo owner, and controls the UAV to hover in front of the cargo owner without stopping the operation of the UAV to reduce its energy consumption. First, the data acquisition module is used to obtain the information of the goods to be delivered, and the The face image of the cargo owner corresponding to the cargo, this face image is a standard image used in the subsequent comparison process, the data acquisition module includes an input unit, which is used to obtain the facial image uploaded by the cargo owner, and calibrate the facial image as For standard images, the upper limit of face images uploaded by consignors is set to 3 to 5. This standard image needs to clearly show the facial features of the human body. Of course, before the delivery of the goods, it is necessary to communicate with the consignee first to ensure The goods can be accurately delivered to the owner. At the same time, it should be clear whether the owner changes the destination. This is not a technical point that needs to be protected in this field. I will not elaborate here. After the drone carries the goods to the designated location , first collect the human body image of the cargo owner, the human body image includes a face image and a joint image, the joint image is the arm and shoulder joints of the cargo owner, and then use the first-level judgment module to judge whether the facial image of the cargo owner is consistent with the standard image, and if the two are consistent In this case, continue to determine the positional relationship between the arm joints and the shoulder joints, and determine the delivery method according to the posture of the owner, so as to decide whether to drop the goods on the ground or to pick up the goods by hovering. When the owner observes the drone , there may be a delay in making the gesture of receiving the goods. In order to ensure the delivery efficiency of the goods, this embodiment also provides a motion capture module to capture the action information of the owner of the goods, and combine the feature analysis module to calculate the movement trend of the owner. Finally, the prediction module is used to estimate the predicted attitude of the cargo owner, and then the predicted attitude is compared with the standard attitude image, and the position of the hovering point of the UAV is adjusted in real time according to the comparison results, so as to complete the delivery of the goods. When the receiving posture is not made, the goods are directly placed on the ground in front of the owner of the goods, and when the clamping state of the goods is released, it is necessary to use the unloading judgment module to determine whether the force of the goods meets the unloading parameters, and the unloading parameter settings is the pressure change parameter between the cargo and the UAV. The pressure change parameter is set according to the weight of the cargo. The value of the pressure change parameter is greater than half of the cargo weight. The purpose is to make the cargo not touch the ground. The inclination also enables the cargo owner to adapt to the weight of the cargo when receiving the cargo, and avoids the phenomenon that the cargo suddenly falls and hurts the cargo owner.

In a preferred embodiment, it also includes a communication module, which is used to remind the owner of the cargo to choose the delivery method;

If the cargo owner chooses to hover to pick up the cargo, the drone will hover in front of the cargo owner;

If the cargo owner chooses to drop the cargo on the ground, the drone will place the cargo on the ground;

If the owner of the goods does not respond, the body image of the owner is recognized through the first-level judgment module, and the delivery method is automatically determined according to the recognition result.

In this embodiment, the communication module can be used to remind the cargo owner to choose the delivery method independently before the delivery of the goods, because there may be fragile or valuable items in the goods delivered by the drone, so communicate with the owner first and verify the delivery method It is very necessary, and when the owner of the goods does not respond, the delivery method is determined independently through the judgment result of the first-level judgment module.

Secondly, the feature extraction module includes a first execution unit and a second execution unit, and the execution priority of the first execution unit is higher than that of the second execution unit;

When the first execution unit executes, it captures the facial image of the cargo owner, and uploads the facial image to the first-level judgment module;

When the second execution unit executes, it captures the joint image of the cargo owner and uploads the joint image to the first-level determination module, wherein the joint image includes shoulder joint features, elbow joint features and wrist joint features.

Among the above, the front end of the UAV is equipped with a binocular camera, which is used to collect the face image and joint image of the owner, and the priority of collecting face images is higher than that of joint images. When the UAV arrives at the specified After the location, the binocular camera will collect the human body image and motion information of the cargo, and upload the human body image to the first-level judgment module and the second-level judgment module to provide data support for whether to lift the restrictions on the goods.

Next, the primary judgment module includes a first judgment unit and a second judgment unit, and the execution priority of the first judgment unit is higher than that of the second judgment unit;

The first judging unit is used to judge whether the face image is consistent with the standard image;

If they are consistent, it is determined that the information of the cargo owner is accurate, and the drone is hovered directly in front of the cargo owner;

If inconsistent, it is determined that the owner information does not match, and the second determination unit is not executed;

The second judging unit is used for judging the positional relationship of the shoulder joint features, the elbow joint features and the wrist joint features;

If the wrist joint feature is above the elbow joint feature and the elbow joint feature is above the shoulder joint feature, it is determined that the owner chooses to hover in the air to pick up the goods;

If the shoulder joint feature is above the elbow joint feature, it is determined that the cargo owner chooses to drop the cargo on the ground.

In this embodiment, after the binocular camera captures the owner's face image, it is immediately uploaded to the first judging unit, and the first judging unit compares the face image with the standard image, and judges whether the comparison result is Consistent, and in the case of the same comparison results, execute the second determination unit, the second determination unit determines the positional relationship between the shoulder joint features, elbow joint features and wrist joint features, and the wrist joint features, elbow joint features and shoulder joint features When the features are arranged in order from high to low, it can be determined that the cargo owner is in the receiving posture, and it is determined that the cargo owner chooses to hover to receive the cargo. At this time, the UAV can carry the cargo and hover in front of the cargo owner. The shoulder joint feature of the cargo owner is located at the elbow. When the joint feature is above, it is determined that the owner chooses to drop the goods on the ground, and the owner’s receiving posture may be delayed. At this time, it is necessary to combine the motion capture module to collect the movement information of the owner, output the movement trend of the owner, and judge the owner’s wrist joint. Whether the feature is rising, if so, it is determined that the owner of the goods chooses to hover to pick up the goods, otherwise, it is determined that the owner of the goods chooses to drop the goods on the ground;

Here, it needs to be explained that after the wrist joint feature, elbow joint feature and shoulder joint feature are determined, a virtual coordinate system is constructed with the shoulder joint feature as the origin, and the edge coordinates of the wrist joint feature, elbow joint feature and shoulder joint feature are determined point, and compare the vertical coordinates of these edge coordinate points.

Secondly, the standard posture image is preset in the secondary judgment module, and compared with it after obtaining the predicted posture;

If the predicted attitude is consistent with the standard attitude image, the UAV hovers at the specified position;

If the predicted attitude is inconsistent with the standard attitude image, a deviation distance is generated, and the hovering position of the UAV is adjusted to the position of the predicted attitude.

In this embodiment, when predicting the arm posture of the owner, the motion information of the owner is first obtained through the motion capture module, which is video information, and after the video information is captured, it is input into the decoder for decoding After processing, athletes can obtain multiple frames of continuous joint images, and obtain the edge coordinates of the shoulder joint features, elbow joint features, and wrist joint features in these joint images, and mark these edge coordinates as the coordinates to be evaluated, and then from the trend The trend analysis function is called in the analysis model, and the coordinates to be evaluated are input into the trend analysis function, and then the change trend value of the coordinates to be evaluated can be obtained, and the coordinates to be evaluated are set as the wrist joint features, elbow joint features and shoulder joint features. The ordinate, where the trend analysis function is: , where, /> Indicates the change trend value of the coordinates to be evaluated, /> Indicates the number of coordinates to evaluate, /> Indicates the number of the coordinate to be evaluated, and does not participate in the actual operation, /> and /> Both represent the coordinates to be evaluated. After the change trend value of the coordinates to be evaluated is obtained, it is judged immediately. If the value of the change trend value of the coordinates to be evaluated is negative, it is determined that the owner is not in the attitude of receiving goods, and the default owner chooses For ground delivery, if the change trend value of the coordinates to be evaluated is positive, it is determined that the owner has a receiving trend, then it is determined that the owner chooses to hover to pick up the goods, and the predicted attitude is calculated. This predicted attitude is the characteristic of the owner's wrist joint, Elbow joint features and shoulder joint features are arranged according to the critical point coordinates from high to low. If the predicted attitude is consistent with the standard image attitude, the drone will not change the hovering point. Otherwise, the wrist joint features in the standard image and the predicted The deviation distance of the wrist joint feature in the posture, and the drone can adjust the hovering point according to the deviation distance.

In a preferred embodiment, a voice broadcast module is also included, and the voice broadcast module is used to play reminder acceptance sentences, such as "please confirm that the goods are in good condition" or "please accept the goods".

As shown in Figure 3, the present invention also provides a logistics drone based on human body posture recognition, which is applied to the above-mentioned human-machine logistics interaction system based on human body posture recognition, including a fuselage 1, and a useful drone is arranged under the fuselage 1. For the clamping mechanism 2 carrying goods, a tension sensor is provided on the clamping mechanism 2, and the tension sensor is used to detect the tension value between the clamping mechanism 2 and the carried goods;

Among them, after the first-level determination module determines the unloading mode, the unloading determination module sets the unloading parameter according to the unloading mode. The unloading parameter is the pressure change value between the clamping mechanism and the loaded cargo. When the force of the holding mechanism 2 satisfies the unloading parameters, the holding mechanism releases the holding of the goods.

In the above, after the loaded cargo falls to the ground or the palm of the cargo owner, the loaded cargo will be supported by the ground or the lifting force of the cargo owner’s palm. At this time, the force of the clamping mechanism 2 is reduced. Correspondingly, the tension sensor The tension value of the pull force will also decrease, and when the reduced value meets the unloading parameter, the clamping mechanism 2 will release the clamping of the loaded goods, and then the owner of the goods can take away the goods.

In addition, the clamping mechanism 2 includes a clamping plate 201 and a two-way telescopic rod 202, two clamping plates 201 are provided, the two clamping plates 201 are arranged below the fuselage 1, and the two-way telescopic rod 202 is fixedly installed on the fuselage 1, and the two ends of the two-way telescopic rod 202 are respectively fixedly connected with the top ends of the two clamping plates 201.

Further, the two clamping plates 201 are both arranged in an L-shape, and an outer box for carrying goods is arranged between the two clamping plates 201, and the two sides of the outer box for carrying goods are provided with corresponding openings for the holding plates 201. snap slot.

Specifically, after the reduction value of the tension sensor satisfies the unloading parameter, the two-way telescopic rod 202 moves, and drives the two clamping plates 201 to move away from the two sides of the outer box carrying the goods. At this time, the outer box carrying the goods The restriction will be lifted, and then the owner of the goods can take away the goods.

It should be noted that, in this document, the terms "comprising", "comprising" or any other variation thereof are intended to cover a non-exclusive inclusion such that a process, apparatus, article or method comprising a set of elements includes not only those elements, It also includes other elements not expressly listed, or elements inherent in the process, apparatus, article, or method. Without further limitations, an element defined by the phrase "comprising a ..." does not preclude the presence of additional same elements in the process, apparatus, article or method comprising the element.

The above is only a preferred embodiment of the present invention, it should be pointed out that for those of ordinary skill in the art, without departing from the principle of the present invention, some improvements and modifications can also be made, and these improvements and modifications should also be It is regarded as the protection scope of the present invention. The structures, devices and operation methods not specifically described and explained in the present invention, unless otherwise specified and limited, shall be implemented according to conventional means in the art.

Claims (7)

1. A human-computer logistics interaction system based on human gesture recognition, including a data acquisition module, a feature extraction module, a first-level judgment module, an unloading judgment module, a motion capture module, a feature analysis module, a prediction module, a second-level judgment module, and an intermediate control module, characterized in that: The data collection module is used to collect cargo information and the face image of the owner, and mark the face image of the owner as a standard image; The feature extraction module is used to extract the human body image of the cargo owner, wherein the human body image includes a face image and a joint image; The feature extraction module includes a first execution unit and a second execution unit, the execution priority of the first execution unit is higher than that of the second execution unit; When the first execution unit executes, it captures the facial image of the cargo owner and uploads the facial image to the primary judgment module; When the second execution unit is executed, it captures the joint image of the consignor and uploads the joint image to the first-level judgment module, wherein the joint image includes shoulder joint features, elbow joint features and wrist joint features; The first-level judgment module is used to identify the human body image of the cargo owner, and determine the unloading method according to the recognition result, wherein the unloading method includes hovering in the air to pick up the goods and ground unloading; The primary judgment module includes a first judgment unit and a second judgment unit, and the execution priority of the first judgment unit is higher than that of the second judgment unit; The first judging unit is used to judge whether the facial image is consistent with a standard image; If they are consistent, it is determined that the information of the cargo owner is accurate, and the drone is hovered directly in front of the cargo owner; If inconsistent, it is determined that the owner information is inconsistent, and the second determination unit is not executed; The second judging unit is used to judge the positional relationship of the shoulder joint features, elbow joint features and wrist joint features; If the wrist joint feature is located above the elbow joint feature and the elbow joint feature is located above the shoulder joint feature, it is determined that the cargo owner chooses to hover in the air to pick up the goods; If the shoulder joint feature is located below the elbow joint feature, it is determined that the owner chooses to drop the goods on the ground; The unloading determination module is used to set the unloading parameters according to the unloading mode, the unloading parameters are set as the pressure change parameters between the cargo and the UAV, and the pressure change parameters are set according to the weight of the cargo, wherein the pressure change The value of the parameter is greater than half of the weight of the goods, and when the unloading parameters are met, the restriction on the goods will be lifted; The motion capture module is used to collect the motion information of the owner, decode the motion information, output multiple frames of continuous images, and mark them as images to be analyzed; The feature analysis module is used to obtain all the images to be analyzed, and input the images to be analyzed into the trend analysis model, and output the movement trend of the owner; The prediction module is used to predict the arm posture of the cargo owner according to the movement trend, and output it as the predicted posture; The secondary judgment module is used to obtain the predicted attitude, and adjust the hovering position of the drone according to the predicted attitude; The central control module is used to send and receive transfer information among the data collection module, feature extraction module, primary judgment module, motion capture module, feature analysis module, prediction module and secondary judgment module. 2. A human-computer logistics interactive system based on human gesture recognition according to claim 1, characterized in that: it also includes a communication module, and the communication module is used to remind the owner of the cargo to select the delivery method; If the cargo owner chooses to hover to pick up the cargo, the drone will hover in front of the cargo owner; If the cargo owner chooses to drop the cargo on the ground, the drone will place the cargo on the ground; If the owner of the goods does not respond, the body image of the owner is recognized by the first-level judgment module, and the delivery method is automatically determined according to the recognition result. 3. A kind of human-computer logistics interaction system based on human body posture recognition according to claim 1, characterized in that: a standard posture image is preset in the secondary judgment module, and is compared with it after obtaining the predicted posture; If the predicted attitude is consistent with the standard attitude image, the UAV hovers at the designated position; If the predicted attitude is inconsistent with the standard attitude image, a deviation distance is generated, and the hovering position of the drone is adjusted to the position of the predicted attitude. 4 . The human-computer logistics interaction system based on human gesture recognition according to claim 1 , further comprising a voice broadcast module, and the voice broadcast module is used to play reminder acceptance sentences. 5. A logistics unmanned aerial vehicle based on human body posture recognition, applied to the human-computer logistics interaction system based on human body posture recognition according to any one of claims 1 to 4, characterized in that: comprising: The fuselage (1), the clamping mechanism (2) for carrying goods is arranged under the fuselage (1), the tension sensor is arranged on the clamping mechanism (2), and the tension sensor is used to detect the clamping The tension value between the mechanism (2) and the cargo it carries; Wherein, after the first-level determination module determines the unloading mode, the unloading determination module sets the unloading parameter according to the unloading mode, and the unloading parameter is the pressure change value between the clamping mechanism and the loaded goods. When the tension sensor detects that the force of the clamping mechanism (2) satisfies the parameters of unloading, the clamping mechanism releases the clamping of the cargo. 6. A logistics drone based on human gesture recognition according to claim 5, characterized in that: the clamping mechanism (2) includes a clamping plate (201) and a two-way telescopic rod (202), the Two clamping plates (201) are provided, and the two clamping plates (201) are arranged under the fuselage (1), and the two-way telescopic rod (202) is fixedly installed on the bottom of the fuselage (1) end, and the two ends of the two-way telescopic rod (202) are respectively fixedly connected to the top ends of the two clamping plates (201). 7. A logistics drone based on human body posture recognition according to claim 6, characterized in that: the two clamping plates (201) are both set in an L shape, and the two clamping plates (201 ) is provided with an outer cargo-carrying box, and both sides of the cargo-carrying outer box are provided with clamping slots corresponding to the clamping plate (201).