A device and method for automatic sampling of living whales at a long distance based on machine vision

Abstract

The present invention discloses a long-distance automatic sampling device and method for living whales based on machine vision. The long-distance automatic sampling device for living whales comprises a bionic drone, on which a sampling system is installed. The sampling system comprises a transmitter and a sampling gun, and the sampling gun comprises a gun body and a hollow sampling gun head. The hollow sampling gun head comprises a accommodating chamber and a boss structure connected to the accommodating chamber, and one-way drainage valves and movable baffles are respectively arranged on both sides of the accommodating chamber, and spring extraction mechanisms are arranged around the outer walls on both sides of the accommodating chamber. When the hollow sampling gun head successfully penetrates into the target organism, the one-way drainage valve is used to form a negative pressure in the accommodating chamber. After the spring extraction mechanism extracts the hollow sampling gun head, the accommodating chamber carries out the sample in the target organism, thereby effectively extracting the biological sample. Under the elastic force of the spring extraction mechanism, the hollow sampling gun head can be quickly extracted after the sampling is completed, thereby minimizing the damage to the organism.

Description

Technical Field

[0001] The invention belongs to the technical field of underwater large-scale biological live body sampling devices, and in particular relates to a long-distance whale live body automatic sampling device and method based on machine vision. Background Art

[0002] Marine mammals are a special group of mammals adapted to aquatic environments, including all species of the orders Cetaceans, Sirenians, and Pinnipeds, and some groups of the order Carnivora. Most of them are at the top of the food chain of the marine ecosystem and are of great significance to the stability of the marine ecosystem. Compared with most other marine animals, whales are scarce in number and live underwater most of the time. Some species even live in the open ocean and deep sea environments, and have the ability to swim quickly. Whales are also protected animals. These characteristics make it difficult to obtain samples of whales, resulting in long-term slow progress in research work that relies on whale tissue samples. Due to the inaccessibility of whales, the long-distance live sampling method has become one of the most suitable non-lethal sampling methods for sampling them. Over the past 40 years, the long-distance live sampling method has been used for sampling more than 40 species of whales, and thousands of whale tissue samples have been obtained, which have been used in hot research areas such as identification of kinship between individuals, feeding ecology, and enrichment of pollutants. It has provided important research materials for research on whale genetics and physiology around the world.

[0003] The deficiencies and reasons of existing technologies: 1. Most existing technologies have a low degree of automation and rely heavily on manual operation. During the operation, in order to avoid unnecessary damage, the operator needs to maintain a relatively close distance with the whale for a long time underwater. On the one hand, the sampling success rate is low, and on the other hand, the operator is easily injured by the whale's violent swinging and slapping after being stimulated. 2. Most existing technologies have a low degree of intelligence and are too dependent on the operator's work experience. They cannot effectively use sensors and image recognition technologies to achieve automatic and precise aiming, and cannot make intelligent adjustments based on real-time environmental changes, resulting in low work efficiency and low sampling success rate. 3. The existing sampling structure has not been designed and optimized for the characteristics of whales, resulting in greater damage to the organisms, which is not conducive to the protection of the marine ecological environment. On the other hand, the sampling head is easy to fall off during the recovery process, resulting in a low sampling success rate. Summary of the invention

[0004] The present invention provides a long-distance live whale automatic sampling device and method based on machine vision, so as to solve the problems of low sampling accuracy, low operation efficiency and poor environmental adaptability of the existing long-distance live whale sampling technology.

[0005] In order to solve the above problems, the present invention proposes the following technical solutions:

[0006] An embodiment of the present invention provides a long-distance automatic sampling device for live whales based on machine vision, comprising a bionic drone (1), wherein a sampling system is installed on the bionic drone (1), wherein the sampling system comprises a transmitter and a sampling gun (4), wherein the transmitter comprises a trigger system (7), an ejection system (8), a sponge shock absorbing mechanism (9) and an electric winding wheel (10); wherein the trigger system (7) is connected to the ejection system (8), wherein the sponge shock absorbing mechanism (9) is arranged between the ejection system (8) and the housing of the bionic drone (1), and wherein the electric winding wheel (10) is installed at the bottom of the front end of the bionic drone (1); wherein the sampling gun (4) comprises a gun body (4-1) and a hollow sampling gun head ( 4-2), the gun body (4-1) is arranged to abut against the ejection system (8), the hollow sampling gun head (4-2) is installed at the end of the gun body (4-1), the hollow sampling gun head (4-2) comprises a receiving chamber (4-3) and a boss structure (4-4) connected to the receiving chamber (4-3), the inner wall of the boss structure (4-4) is a continuous internal thread structure (4-8), a one-way drainage valve (4-5) and a movable baffle (4-6) are respectively arranged on both sides of the receiving chamber (4-3), and a spring extraction mechanism (4-7) is arranged around the outer walls of both sides of the receiving chamber (4-3); the other end of the gun body (4-1) is connected to the electric fishing reel (10) through a fishing line (12);

[0007] The gun body (4-1) provides kinetic energy to the hollow sampling gun head (4-2) through the emitter. The hollow sampling gun head (4-2) is used to penetrate a target organism. When the hollow sampling gun head (4-2) successfully penetrates into the target organism, the one-way drainage valve (4-5) is used to form a negative pressure in the accommodating chamber (4-3). After the spring extraction mechanism (4-7) extracts the hollow sampling gun head (4-2), the accommodating chamber (4-3) carries out the sample in the target organism. The continuous internal thread structure (4-8) on the inner wall of the boss structure (4-4) is used to effectively prevent sample leakage.

[0008] In an optional embodiment of the present invention, the bionic drone (1) is a fish-like structure having a fish head portion (1-1), a three-section body structure (1-2), a tail fin (1-3), and left and right balance fins (1-4), and each section of the three-section body structure (1-2) is driven by a servo motor (2).

[0009] In an optional embodiment of the present invention, the fish head portion (1-1) is provided with a posture sensor (3) for real-time monitoring of motion state information of the underwater bionic drone, wherein the motion state information includes turning and flipping phenomena.

[0010] In an optional embodiment of the present invention, the one-way drain valve (4-5) comprises a drain port and an O-shaped rubber ring, the drain port is arranged on the inner wall of the accommodating cavity (4-3), and the O-shaped rubber ring is sleeved on the outside of the drain port.

[0011] In an optional embodiment of the present invention, the movable baffle (4-6) is connected to the spring extraction mechanism (4-7), and when the boss structure (4-4) penetrates into a target organism, the spring extraction mechanism (4-7) is compressed, the movable baffle (4-6) moves backward, and the accommodating chamber (4-3) is opened; when the spring extraction mechanism (4-7) restores its length, the movable baffle (4-6) moves forward, and the accommodating chamber (4-3) is closed.

[0012] In an optional embodiment of the present invention, the trigger system (7) comprises a motor (7-1) and a trigger (7-2), one end of the trigger (7-2) is connected to the output shaft of the motor (7-1), and the other end is clamped under the ejection system (8).

[0013] In an optional embodiment of the present invention, the front end bottom of the bionic drone (1) is provided with an environment perception system (11), and the environment perception system (11) performs real-time monitoring and identification of the surrounding environment and target organisms through high-precision sensors and a YOLOv10 target recognition algorithm module.

[0014] In an optional embodiment of the present invention, the bionic drone (1) is equipped with a pressure sensor (5) and a sonar sensor (6), wherein the pressure sensor (5) is used to obtain the pressure of seawater, and the sonar sensor (6) is used to detect and identify the outline of a target organism in the water.

[0015] The embodiment of the present invention provides a method for automatic sampling of living whales from a distance based on machine vision, which is implemented by a device for automatic sampling of living whales from a distance based on machine vision according to the above embodiment, and includes the following steps:

[0016] Step 1, when the bionic drone (1) reaches the predetermined target area, the environmental perception system (11) is activated, and the surrounding environment and target organisms are monitored and identified in real time through high-precision sensors and the YOLOv10 target recognition algorithm module; after determining the specific location and motion state of the target organism, the bionic drone (1) gradually adjusts its own posture and navigation path under the coordination of the visual servo and PID control system, and accurately approaches the target until it enters the preset effective shooting range;

[0017] Step 2, after entering the shooting range, the trigger system (7) is started, and the rotary motor (7-1) drives the rear end of the trigger (7-2) to be pressed down and the front end to be lifted, so that the trigger (7-2) structure is unlocked; at this time, the sampling gun (4) is quickly fired under the action of the ejection system (8), and the hollow sampling gun head (4-2) accurately penetrates into the target organism; at the moment when the hollow sampling gun head (4-2) penetrates into the target, the water flow in the containing chamber (4-3) is discharged through the one-way drain valve (4-5), forming a negative pressure environment, thereby effectively sucking the sample in the target organism. At the same time, the spring extraction mechanism (4-7) converts part of the kinetic energy into elastic potential energy when the boss structure (4-4) penetrates, and releases this part of the energy after the sampling is completed, and quickly extracts the boss structure (4-4) from the organism to minimize the damage to the target organism; during the extraction process, the negative pressure suction effect sucks the sample in the organism into the containing cavity (4-3), and the continuous internal thread structure (4-8) ensures the stable storage of the sample to prevent the sample from leaking;

[0018] Step 3, after the sampling is completed, the environmental perception system (11) confirms whether the sampling process is successful by real-time monitoring of the status of the sampling gun (4); when it is detected that the boss structure (4-4) has been completely pulled out of the target body and the sample is collected, the environmental perception system (11) automatically starts the electric reel (10) to drive the fishing line (12) connected to the sampling gun (4) to perform a recovery operation; the electric reel (10) drives the fishing line (12) to safely recover the sampling gun (4) to the bionic drone (1) platform, thereby ensuring the safe recovery and preservation of the sample.

[0019] Compared with the prior art, the embodiment of the present invention provides a long-distance automatic sampling device and method for living whales based on machine vision, which has the following beneficial effects:

[0020] (1) The present invention has an efficient and autonomous function. Compared with traditional underwater sampling, which usually relies on manual divers to perform target identification and sampling, the present invention can autonomously complete target identification, posture adjustment and sampling tasks without manual intervention by introducing automation technology. Based on the yolov10 target recognition algorithm module, the present invention can process and analyze underwater image data in real time, accurately identify target organisms, greatly reduce sampling time, and increase the frequency of data collection. After the target is determined, the system can quickly adjust the posture and use the PID control algorithm to ensure that the drone enters the effective sampling range. Through this efficient and autonomous method, not only the operating efficiency is improved, but also the errors caused by human factors are reduced, ensuring the accuracy of sampling.

[0021] (2) The present invention has higher collection efficiency and operation economy. The present invention designs a small-caliber hollow sampling head, which is equipped with a spring device to control the penetration depth, ensuring that the impact on the target organism during the sampling process is minimized. When the sampling head successfully penetrates the target organism, the suction device forms a negative pressure in the hollow gun head through the one-way drain valve, thereby effectively extracting the biological sample. With the help of the spring, the sampling head can be quickly pulled out after the sampling is completed, minimizing damage to the organism. In addition, the internal thread design effectively prevents sample leakage and ensures the integrity of the sampling process. This low-damage sampling technology not only improves the quality and reliability of sample collection, but also provides a safer and more scientific technical means for ecological research.

[0022] (3) The flexibility and versatility of the modular design of the present invention. The present invention adopts a highly modular design concept, such as the environmental perception system, pressure sensor and sonar sensor. The bionic drone can flexibly replace functional modules according to environmental requirements to enhance its adaptability. At the same time, the sampling head can be flexibly configured according to different biological species and sampling requirements, thereby minimizing damage to the organisms. At the same time, the modular design also facilitates the updating and maintenance of technology. Operators can improve efficiency and extend the service life of equipment by integrating new sensors. In addition, the modular structure can also promote interdisciplinary cooperation, enabling different research teams to customize and share resources, thereby promoting innovation and development in research. This gives the system greater flexibility and versatility, giving it a wider application potential in ecological monitoring and biological research. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] In order to more clearly illustrate the technical solutions in the embodiments or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0024] Figure 1 A front view of a long-distance automatic sampling device for living whales based on machine vision provided in an embodiment of the present application.

[0025] Figure 2 A right view of a long-distance automatic sampling device for living whales based on machine vision provided in an embodiment of the present application.

[0026] Figure 3 for Figure 2 Schematic diagram of the AA section.

[0027] Figure 4 for Figure 3 The enlarged schematic diagram of point B in FIG.

[0028] Figure 5 Schematic diagram of a sampling gun of a long-distance automatic sampling device for living whales based on machine vision provided in an embodiment of the present application, which sucks samples from a target organism.

[0029] Figure 6 A schematic diagram of a sampling gun of a long-distance automatic sampling device for living whales based on machine vision provided in an embodiment of the present application being detached from a target organism.

[0030] Figure 7 A schematic diagram of the working principles of an environmental perception system, a pressure sensor, and a sonar sensor provided in an embodiment of the present application.

[0031] Figure 8 A schematic diagram of a method for extracting a sample from a target organism provided in an embodiment of the present application. DETAILED DESCRIPTION

[0032] The technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present application.

[0033] like Figure 1 to Figure 6 As shown, an embodiment of the present invention provides a long-distance automatic sampling device for living whales based on machine vision, including a bionic drone 1, on which a sampling system is installed, the sampling system includes a launcher and a sampling gun 4, the launcher includes a trigger system 7, an ejection system 8, a sponge shock absorbing mechanism 9 and an electric reel 10; the trigger system 7 is connected to the ejection system 8, the sponge shock absorbing mechanism 9 is arranged between the ejection system 8 and the shell of the bionic drone 1, and the electric reel 10 is installed at the bottom of the front end of the bionic drone 1; the trigger system 7 and the ejection system 8 jointly provide kinetic energy for the sampling gun 4; the sponge shock absorbing mechanism 9 plays a buffering role to reduce the damage to the sampling system after launching; the electric reel 10 recovers the sampling gun 4 by winding the fishing line 12.

[0034] The sampling gun 4 includes a gun body 4-1 and a hollow sampling gun head 4-2. The gun body 4-1 is set in conflict with the ejection system 8. The hollow sampling gun head 4-2 is installed at the end of the gun body 4-1. The hollow sampling gun head 4-2 includes a accommodating chamber 4-3 and a boss structure 4-4 connected to the accommodating chamber 4-3. The inner wall of the boss structure 4-4 is a continuous internal thread structure 4-8. A one-way drain valve 4-5 and a movable baffle 4-6 are respectively arranged on both sides of the accommodating chamber 4-3. A spring pull-out mechanism 4-7 is arranged around the outer walls of both sides of the accommodating chamber 4-3; the other end of the gun body 4-1 is connected to the electric reel 10 through a fishing line 12. The hollow sampling gun head 4-2 uses a small-caliber hollow sampling head to penetrate the target organism. The hollow sampling gun head 4-2 and the gun body 4-1 are connected by threads, and the spring pull-out mechanism 4-7 is fixed on the hollow sampling gun head 4-2 and controls the penetration depth of the movable baffle 4-6.

[0035] like Figure 5 and Figure 6 As shown, the gun body 4-1 provides kinetic energy to the hollow sampling gun head 4-2 through the emitter, and the hollow sampling gun head 4-2 is used to penetrate the target organism. When the hollow sampling gun head 4-2 successfully penetrates into the target organism 13, the one-way drain valve 4-5 is used to form a negative pressure in the accommodating chamber 4-3, so that after the spring pulling mechanism 4-7 pulls out the hollow sampling gun head 4-2, the accommodating chamber 4-3 can carry out the sample in the target organism, and the continuous internal thread structure 4-8 on the inner wall of the boss structure 4-4 is used to effectively prevent sample leakage.

[0036] like Figure 1 and Figure 2 As shown, the bionic drone 1 is a fish-like structure having a fish head portion 1-1, a three-section body structure 1-2, a tail fin 1-3, and left and right balancing fins 1-4. Each section of the three-section body structure 1-2 is driven by a servo motor 2, and the left and right balancing fins 1-4 are used to keep the swimming stable.

[0037] The fish head 1-1 is provided with a posture sensor 3, which is used to monitor the motion state information of the underwater bionic drone in real time to ensure stable motion. The motion state information includes turning and flipping phenomena. The bionic drone 1 is equipped with a pressure sensor 5 and a sonar sensor 6. The pressure sensor 5 is used to obtain the pressure of seawater, and the sonar sensor 6 is used to detect and identify the outline of the target organism in the water.

[0038] like Figure 3 and Figure 4As shown, the one-way drain valve 4-5 includes a drain port and an O-shaped rubber ring. The drain port is arranged on the inner wall of the accommodating chamber 4-3, and the O-shaped rubber ring is arranged outside the drain port. The movable baffle 4-6 is connected to the spring extraction mechanism 4-7. When the boss structure 4-4 penetrates the target organism, the spring extraction mechanism 4-7 is compressed, the movable baffle 4-6 moves backward, and the accommodating chamber 4-3 is opened; the spring extraction mechanism 4-7 restores its length, the movable baffle 4-6 moves forward, and the accommodating chamber 4-3 is closed. The trigger system 7 includes a motor 7-1 and a trigger 7-2. One end of the trigger 7-2 is connected to the output shaft of the motor 7-1, and the other end is clamped under the ejection system 8. An environmental perception system 11 is arranged at the bottom of the front end of the bionic drone 1. The environmental perception system 11 monitors and identifies the surrounding environment and target organisms in real time through high-precision sensors and the yolov10 target recognition algorithm module, providing a field of view for target recognition.

[0039] The present invention uses bionics to reduce the vigilance of organisms. Compared with manual operation, it can get closer to the target and achieve more accurate aiming. The electric reel is connected to the gun body through the fishing line, which is convenient for recovery after sampling. In order to achieve low-damage sampling, a small-caliber hollow sampling head is used, and the penetration depth is controlled by a spring pull-out mechanism, and it is pulled out in time, thereby minimizing damage to the organism and ensuring the safety and effectiveness of the sampling process.

[0040] The embodiment of the present invention provides a method for automatic sampling of living whales from a distance based on machine vision, which is implemented by a device for automatic sampling of living whales from a distance based on machine vision in the above embodiment, and includes the following steps:

[0041] Step 1, when the bionic drone 1 reaches the predetermined target area, the environmental perception system 11 is started, and the surrounding environment and target organisms are monitored and identified in real time through high-precision sensors and the yolov10 target recognition algorithm module; after determining the specific location and motion state of the target organism, the bionic drone 1 gradually adjusts its own posture and navigation path under the coordination of the visual servo and PID control system, and approaches the target accurately until it enters the preset effective shooting range;

[0042] Step 2, after entering the shooting range, the trigger system 7 is started, and the rotary motor 7-1 drives the rear end of the trigger 7-2 to press down and the front end to lift, so that the trigger 7-2 structure is unlocked; at this time, the sampling gun 4 is quickly fired under the action of the ejection system 8, and the hollow sampling gun head 4-2 accurately penetrates into the target organism; at the moment when the hollow sampling gun head 4-2 penetrates into the target, the water flow in the accommodating chamber 4-3 is discharged through the one-way drain valve 4-5, forming a negative pressure environment, thereby effectively sucking the sample in the target organism 13. At the same time, the spring extraction mechanism 4-7 converts part of the kinetic energy into elastic potential energy when the boss structure 4-4 penetrates, and releases this part of energy after the sampling is completed, and quickly pulls the boss structure 4-4 out of the organism to minimize damage to the target organism; during the extraction process, the negative pressure suction effect sucks the sample in the organism into the accommodating chamber 4-3, and at the same time the continuous internal thread structure 4-8 ensures stable storage of the sample to prevent sample leakage;

[0043] Step 3, after the sampling is completed, the environmental perception system 11 confirms whether the sampling process is successful by real-time monitoring of the status of the sampling gun 4; when it is detected that the boss structure 4-4 has been completely pulled out of the target body and the sample is collected, the environmental perception system 11 automatically starts the electric reel 10, driving the fishing line 12 connected to the sampling gun 4 for recovery; the electric reel 10 drives the fishing line 12 to safely recover the sampling gun 4 to the bionic drone 1 platform to ensure the safe recovery and preservation of the sample.

[0044] like Figure 7 As shown, step 1 adopts the yolov10 target detection algorithm module based on convolutional neural network, combines deep learning technology and underwater image processing technology, and performs feature extraction, target positioning, target recognition, sonar ranging, and deep-sea environment according to the image information collected by the underwater camera, and marks it with a border line. It combines bounding box regression and non-maximum suppression and other technologies to improve the target detection accuracy, ensure stable recognition effect in complex underwater environments, and remotely transmit underwater image information to operators on the water surface in real time to control the bionic drone to perform tracking, attitude adjustment, trigger firing and other actions.

[0045] like Figure 8 As shown in the figure, step 2 uses the PID control strategy to accurately control the movement of the bionic drone so that it approaches the target in the optimal path. During the approach process, the visual servo control technology adjusts the attitude of the bionic drone in real time to ensure accurate alignment. After entering the effective shooting range, the microcontroller automatically triggers the trigger of the sampling gun to perform the launch task. After the sampling is completed, the electric reel recycles the sampling head through the fishing line, and monitors the recovery status based on visual recognition to ensure safety. During the execution of the task, all sensor and image data are transmitted to the ground station in real time for monitoring, and data recording and post-analysis are performed at the same time.

[0046] The present invention provides a long-distance automatic sampling device and method for living whales based on machine vision, which has the following innovative features:

[0047] Innovation 1: Controlling the appearance of the bionic drone: The present invention is based on the principle of bionics, combining the design of the tail fin and the balance fin, so that the device is highly similar to fish in appearance, and through the articulated structure and multiple independent servo motor compartments, it simulates the movement characteristics of fish. Not only is it similar to fish in static appearance, but it can also show a swimming pattern similar to that of biological fish in dynamic movement. This camouflage can reduce the vigilance of whale creatures, thereby improving the success rate of sampling. At the same time, the functional modules are cleverly embedded in the three-section body structure, which does not destroy the camouflage and rapidity of the overall device, but also ensures safe entry into the effective sampling distance, thereby improving the success rate of sampling. Innovation 2: Low-damage sampling technology: In order to overcome the shortcomings of existing live whale sampling that easily damages the target organism, the present invention adopts a small-caliber hollow sampling head, which is designed to minimize the invasiveness to the organism, thereby reducing damage to the organism. A boss structure and a spring are installed at the front end of the sampling head. The penetration depth of the sampling head can be controlled by adjusting the position of the boss structure so that it can be pulled out promptly and smoothly after reaching a predetermined depth. This can not only effectively reduce damage to marine organisms, but also make adaptive adjustments to different organisms, thereby improving the safety and flexibility of the sampling process.

[0048] Innovation point 3: Negative pressure suction of genetic material: In order to reduce the wound wound and increase the amount of biological samples, the present invention forms a one-way water valve through the drain port and the O-shaped rubber ring. When the sampling head penetrates into the organism, the water in the gun is discharged through the one-way valve to form a negative pressure in the hollow gun head, so that after the spring mechanism pulls out the gun head, the sample in the target organism is carried to the maximum extent, and the internal thread structure of the sampling head effectively prevents the leakage of the sample, greatly improving the efficiency of sample collection. Innovation point 4: Intelligent underwater automatic sampling system: In order to automate the operation requirements and lower the operation threshold, the present invention combines the target recognition, PID motion control and visual servo control of the yolov10 target detection algorithm module to realize the fully automated sampling process of the drone - the bionic drone can autonomously identify the target, accurately adjust the posture and enter the effective shooting range, the microcontroller automatically triggers the sampling gun, and after completing the low-damage sampling, the sampling head is efficiently recovered by the electric reel and fishing line, and the recovery status is monitored by visual recognition, which significantly improves the intelligence, accuracy and sampling efficiency of underwater sampling.

[0049] In summary, although the present invention has been disclosed as above in terms of preferred embodiments, the above preferred embodiments are not intended to limit the present invention. A person skilled in the art may make various changes and modifications without departing from the spirit and scope of the present invention. Therefore, the scope of protection of the present invention shall be based on the scope defined in the claims.

Claims

1. A long-distance automatic sampling device for living whales based on machine vision, characterized in that: The bionic drone (1) comprises a sampling system installed on the bionic drone (1), the sampling system comprising a transmitter and a sampling gun (4), the transmitter comprising a trigger system (7), an ejection system (8), a sponge shock absorbing mechanism (9) and an electric winding wheel (10); the trigger system (7) is connected to the ejection system (8), the sponge shock absorbing mechanism (9) is arranged between the ejection system (8) and the housing of the bionic drone (1), and the electric winding wheel (10) is installed at the bottom of the front end of the bionic drone (1); the sampling gun (4) comprises a gun body (4-1) and a hollow sampling gun head (4-2), the gun body (4-1) and the sampling gun head (4-2) are connected to the sampling ... The ejection system (8) is arranged in a resistive manner, the hollow sampling gun head (4-2) is mounted on the end of the gun body (4-1), the hollow sampling gun head (4-2) comprises a receiving chamber (4-3) and a boss structure (4-4) connected to the receiving chamber (4-3), the inner wall of the boss structure (4-4) is a continuous internal thread structure (4-8), a one-way drainage valve (4-5) and a movable baffle (4-6) are respectively arranged on both sides of the receiving chamber (4-3), and a spring extraction mechanism (4-7) is arranged around the outer walls of both sides of the receiving chamber (4-3); the other end of the gun body (4-1) is connected to the electric fishing reel (10) through a fishing line (12); The gun body (4-1) provides kinetic energy to the hollow sampling gun head (4-2) through the emitter. The hollow sampling gun head (4-2) is used to penetrate a target organism. When the hollow sampling gun head (4-2) successfully penetrates into the target organism, the one-way drainage valve (4-5) is used to form a negative pressure in the accommodating chamber (4-3). After the spring extraction mechanism (4-7) extracts the hollow sampling gun head (4-2), the accommodating chamber (4-3) carries out the sample in the target organism. The continuous internal thread structure (4-8) on the inner wall of the boss structure (4-4) is used to effectively prevent the sample from leaking. The movable baffle (4-6) is connected to the spring extraction mechanism (4-7); when the boss structure (4-4) penetrates the target organism, the spring extraction mechanism (4-7) is compressed, the movable baffle (4-6) moves backward, and the accommodating chamber (4-3) is opened; the spring extraction mechanism (4-7) restores its length, the movable baffle (4-6) moves forward, and the accommodating chamber (4-3) is closed.

2. The long-distance automatic sampling device for living whales based on machine vision according to claim 1 is characterized in that: The bionic drone (1) is a fish-like structure having a fish head portion (1-1), a three-section body structure (1-2), a tail fin (1-3) and left and right balancing fins (1-4), and each section of the three-section body structure (1-2) is driven by a servo motor (2).

3. The long-distance automatic sampling device for living whales based on machine vision according to claim 2 is characterized in that: The fish head part (1-1) is provided with a posture sensor (3) for real-time monitoring of the motion state information of the underwater bionic drone, wherein the motion state information includes turning and flipping phenomena.

4. The long-distance automatic sampling device for living whales based on machine vision according to claim 1 is characterized in that: The one-way drainage valve (4-5) comprises a drainage port and an O-shaped rubber ring, the drainage port being arranged on the inner wall of the accommodating cavity (4-3), and the O-shaped rubber ring being sleeved on the outside of the drainage port to ensure a one-way drainage function.

5. The long-distance automatic sampling device for living whales based on machine vision according to claim 1 is characterized in that: The trigger system (7) comprises a motor (7-1) and a trigger (7-2); one end of the trigger (7-2) is connected to the output shaft of the motor (7-1), and the other end is clamped under the ejection system (8).

6. The long-distance automatic sampling device for living whales based on machine vision according to claim 1 is characterized in that: The front bottom of the bionic drone (1) is provided with an environment perception system (11), and the environment perception system (11) performs real-time monitoring and identification of the surrounding environment and target organisms through high-precision sensors and a YOLOv10 target identification algorithm module.

7. The long-distance automatic sampling device for living whales based on machine vision according to claim 1 is characterized in that: The bionic drone (1) is also provided with a pressure sensor (5) and a sonar sensor (6); the pressure sensor (5) is used to obtain the pressure of seawater, and the sonar sensor (6) is used to detect and identify the outline of a target organism in the water.

8. A method for automatic sampling of living whales from a distance based on machine vision, which is implemented by the automatic sampling device for living whales from a distance based on machine vision according to any one of claims 1 to 7, characterized in that: The following steps are involved: Step 1, when the bionic drone (1) reaches the predetermined target area, the environmental perception system (11) is activated, and the surrounding environment and target organisms are monitored and identified in real time through high-precision sensors and the YOLOv10 target recognition algorithm module; after determining the specific location and motion state of the target organism, the bionic drone (1) gradually adjusts its own posture and navigation path under the coordination of the visual servo and PID control system, and accurately approaches the target until it enters the preset effective shooting range; Step 2, after entering the shooting range, the trigger system (7) is started, and the rotary motor (7-1) drives the rear end of the trigger (7-2) to be pressed down and the front end to be lifted, so that the trigger (7-2) structure is unlocked; at this time, the sampling gun (4) is quickly fired under the action of the ejection system (8), and the hollow sampling gun head (4-2) accurately penetrates into the target organism; when the hollow sampling gun head (4-2) penetrates into the target, the water in the accommodating chamber (4-3) is discharged through the one-way drain valve (4-5), forming a negative pressure environment, so that when the sampling gun is pulled out The sample in the target organism is effectively sucked out; at the same time, the spring extraction mechanism (4-7) converts part of the kinetic energy into elastic potential energy when the boss structure (4-4) penetrates, and releases this part of energy after the sampling is completed, and quickly extracts the boss structure (4-4) from the organism to minimize the damage to the target organism; during the extraction process, the negative pressure suction effect sucks the sample in the organism into the containing cavity (4-3), and the continuous internal thread structure (4-8) ensures the stable storage of the sample and prevents the sample from leaking; Step 3, after the sampling is completed, the environmental perception system (11) confirms whether the sampling process is successful by real-time monitoring of the status of the sampling gun (4); when it is detected that the boss structure (4-4) has been completely pulled out of the target body and the sample is collected, the environmental perception system (11) automatically starts the electric reel (10) to drive the fishing line (12) connected to the sampling gun (4) to perform a recovery operation; the electric reel (10) drives the fishing line (12) to safely recover the sampling gun (4) to the bionic drone (1) platform, thereby ensuring the safe recovery and preservation of the sample.

Images