Method for monitoring motion behaviors of estuary offshore fishes based on unmanned aerial vehicle

Abstract

According to the method for efficiently monitoring the motion behaviors of the fishes near the estuary shore based on the unmanned aerial vehicle, by combining the unmanned aerial vehicle high-resolution image technology, the video monitoring technology and the fish motion behavior analysis data processing algorithm, efficient and accurate monitoring of the motion behaviors and distribution of the fishes in the estuary area is achieved. Through data acquisition and accurate motion trail analysis, the method not only improves the monitoring efficiency, but also provides accurate data support for researching fish behaviors, is helpful to better understand the motion migration mode of fish in the estuary area, assesses the habitat quality, and improves the fish habitat quality. And an activity mode of influence of environment change on fish behaviors is monitored, so that a scientific basis is provided for ecological environment protection and fishery sustainable management measures in the near-shore of the estuary.

Description

Technical Field

[0001] The present invention relates to the technical field of ecological monitoring, and particularly relates to a method for monitoring the movement behavior of estuarine and nearshore fish based on an unmanned aerial vehicle (UAV). Background Art

[0002] As the intersection of marine and freshwater ecosystems, the estuarine area is an important habitat and breeding ground for many fish species. Estuarine fishery populations migrate and aggregate along the key habitats of the estuarine and nearshore areas, carrying out important behaviors such as reproduction, foraging, and evasion. The degradation and loss of key habitats for estuarine and nearshore fish affect the aggregation location of fish and their ability to utilize resources and resist predation risks in the migration area, thereby affecting the dispersion and colonization patterns of fish populations. Therefore, by means of biokinetics, monitoring and analyzing the movement behaviors and ranges of estuarine and nearshore fishery species at different life history stages helps to evaluate the response of fish species communities to changes in the functions of estuarine and offshore fishery ecosystems.

[0003] Traditional methods for monitoring estuarine fish mainly include fixed-point surveys by vessels, which have several limitations. These methods are costly and require a large amount of human and material resources. In terms of data collection, these traditional methods face problems of limited coverage and low efficiency, making it difficult to achieve comprehensive monitoring of vast estuarine areas. In addition, traditional methods cannot achieve real-time tracking of fish movement, so it is very difficult to capture their behavioral events. With the development of UAV technology, we can explore a new monitoring method to overcome these limitations and provide more efficient and accurate monitoring data for estuarine fish.

[0004] As an emerging monitoring means, the low-altitude remote sensing technology of unmanned aerial vehicles (UAVs) has developed rapidly in recent years. UAV monitoring technology can effectively overcome the defects of traditional estuarine fish monitoring methods, such as high cost, difficult data collection, low spatio-temporal resolution, and interference with the environment. The high-resolution cameras and multi-spectral sensors carried by UAVs can capture the biological characteristics in the water body, providing clear images, which helps to identify and track fish populations. UAVs can be quickly deployed to rapidly cover vast estuarine areas, greatly improving the efficiency of monitoring work and making it possible to monitor large areas in a short time. Although there have been studies using UAVs to monitor the movement trajectories of terrestrial animals, for the monitoring of estuarine and nearshore fish schools, the application of UAV technology is still in its infancy, mainly limited to the research on the anadromous migration behaviors of large marine mammals and large migratory species such as salmon. The development of UAV technology can provide a new and effective tool for the monitoring of estuarine and nearshore fish schools, and is expected to play an important role in future ecological research and fishery management. Summary of the Invention

[0005] The object of the present invention is to provide a method for monitoring the movement behavior of estuarine and nearshore fish based on unmanned aerial vehicles (UAVs). By combining UAV high-resolution imaging technology, video monitoring technology, and advanced data processing algorithms for fish movement behavior analysis, this method realizes efficient and accurate monitoring of fish movement and distribution in the estuarine area, providing a new technical means for ecological protection and fishery management.

[0006] To achieve the above object, the specific technical solutions adopted by the present invention are as follows:

[0007] The method for monitoring the movement behavior of estuarine and nearshore fish based on UAVs provided by the present invention includes the following steps:

[0008] A. Define the monitoring area and applicable fish populations

[0009] Select the estuarine and nearshore area as the research and monitoring area, and the key areas include tidal creeks, shoals, and areas with rich aquatic plants; monitor the applicable fish populations, which are the dominant species of estuarine fish that are suitable for UAV observation at an altitude of 5-20 meters, often move in the water surface or nearshore shallow water areas, have aggregating behavior, and have visible body lengths, such as Clupeidae, Mugil cephalus, Liza haematocheila, Engraulis japonicus, Sparidae, etc.

[0010] B. Data collection

[0011] Select a UAV with high-resolution imaging capabilities as the shooting tool; according to the life history cycle and tidal cycle of the target fish population, select the shooting time period: the ideal time to monitor estuarine fish is in spring and autumn, especially from March to mid-May and from September to November. During these periods, fish are active, and it is the season when fish come to the estuary to forage or breed.

[0012] Based on the weather conditions, conduct shooting during the rising tide on a sunny day. The water surface is relatively calm and not turbid during the rising tide, which is conducive to UAV observation. The shooting frequency is determined according to the activity habits of the fish and the tidal cycle. Observe the activity trend of the fish group at the beginning of the tidal cycle, and arrange corresponding tracking shooting frequencies according to different activity trend groups.

[0013] During shooting, after searching for the target fish group, lower the flight altitude and maintain a certain flight speed to ensure that all parts of the target fish's body are clearly visible in the field of view; during a single shooting process, it is necessary to arrange fixed-point shooting and tracking mobile shooting at the same time, and ensure that each type of shooting captures key behaviors and covers all aspects of fish activities.

[0014] Specifically, the drone initially departs from the coastline at a height of approximately 30 meters and flies about 200 meters to maximize the search area. After observing the fish school, the drone descends to a height of less than 15 meters and maintains a flight speed of 1 - 3 m / s to observe the fish activities more meticulously, ensuring that all parts of the animal's body are clearly visible in the field of view. The camera is orthogonal to the water surface to minimize distortion and ensure the comparability of all videos.

[0015] Fixed-point shooting is carried out every 5 - 10 minutes to completely record each active group. During a single shooting process, fixed-point shooting and tracking mobile shooting need to be arranged simultaneously, and the duration of each shooting should be maintained at 2 - 3 minutes to ensure capturing key behaviors and covering all aspects of fish activities. Multiple repeated shootings of each fish movement behavior can be carried out within a tidal cycle, focusing on observing the behavioral changes of fish at low tide, the initial stage of rising tide, and high tide, providing sufficient samples for later image processing and data analysis.

[0016] C. Classification analysis of fish behavior and image data processing

[0017] Import the recorded videos into a video analysis tool (Kinovea (v.1.3) software) with point-tracking function. Through comprehensively browsing all video contents, visually conduct a preliminary screening and classification of fish swimming behaviors, that is, when observing sudden changes in fish reactions or behaviors, classify them into one category. According to this classification, select appropriate time intervals of video materials and clip them.

[0018] Specifically, through visual preliminary screening and classification, fish reactions or behaviors are divided into three categories: exploration, rest, and foraging. According to this classification, select appropriate time intervals of video materials, requiring that the time intervals of video materials can reflect each clear and complete fish movement behavior. Preferably, clip them into multiple segments of 20 to 60 seconds for detailed kinematic analysis.

[0019] D. Reconstruction of fish population movement trajectories and discrimination of behavior classification

[0020] Use the point-tracking function to mark the positions of fish in the videos, generating a continuous path that can reflect the fish movement trajectories. By analyzing the time, distance, and speed of these trajectories, analyze and quantify the behavior categories of fish. During the process of marking and tracking the selected fish or fish group, correct the movement trajectories of each frame with the actual movement trajectories at any time to ensure the accuracy of the data. Specifically:

[0021] (1) Use the point tracking function of Kinovea to generate a continuous path that can reflect the movement trajectory of fish. The method is as follows: In the first frame, right-click to mark the position of the tracking object required, click and select "Tracking Path", then continue to right-click and select "Configuration", click the arrow next to "Measurement", then select "Velocity" and click "Apply"; if multiple fish need to be tracked, repeat the above steps; finally, use the "Play" button to move the video clip forward, so that Kinovea can automatically track the selected object in the rest of the clip; after playing is completed, right-click and select "Stop Tracking", and the movement trajectories of all marked fish are generated.

[0022] (2) The method for analyzing and quantifying the behavior categories of fish is as follows: When the fish moves in a medium-speed straight-line trajectory and the vertical movement distance increases linearly, it represents the exploration form of the fish; when the fish moves quickly with local reversals and the vertical movement distance remains basically unchanged, it represents the resting behavior of the fish; when the fish moves slowly in a straight line and then with local reversals, the vertical movement distance increases linearly and then remains basically unchanged, it represents the foraging behavior of the fish.

[0023] Preferably, the medium speed fluctuates in the range of 0 - 240 px / s, the fast speed fluctuates in the range of 0 - 320 px / s, and the slow speed fluctuates in the range of 0 - 100 px / s.

[0024] The technical effects of the technical solution of the present invention are as follows:

[0025] The present invention provides an efficient monitoring method for the movement behavior of fish near the estuary and shore based on drones. This method combines drone high-resolution imaging technology, video monitoring technology, and fish movement behavior analysis data processing algorithms to achieve efficient and accurate monitoring of the movement behavior and distribution of fish in the estuary area. The innovation of the present invention lies in through data collection and precise movement trajectory analysis. This method not only improves the monitoring efficiency, but also provides precise data support for studying fish behavior, helps to better understand the movement and migration patterns of fish in the estuary area, evaluate the habitat quality, and monitor the activity patterns of the impact of environmental changes on fish behavior, providing a scientific basis for the ecological environment protection and fishery sustainable management measures in the estuary and shore. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 It shows the process of monitoring the movement behavior trajectory (a - c), movement displacement (d - f), and movement speed (g - i) of Liza haematocheila near the estuary and shore using the method of the present invention. Among them, a, d, g show the exploration behavior trajectory, displacement, and speed of Liza haematocheila; b, e, h show the resting behavior trajectory, displacement, and speed; c, f, i show the foraging behavior trajectory, displacement, and speed. DETAILED DESCRIPTION OF THE INVENTION

[0027] The following examples and experimental examples further illustrate the present invention and should not be construed as limiting the present invention. The examples do not include detailed descriptions of traditional methods and structures, such methods being well known to those of ordinary skill in the art and having been described in many publications.

[0028] Unless otherwise defined, all professional and scientific terms used herein have the same meaning as those familiar to those skilled in the art. In addition, any methods and structures similar or equivalent to the described content can be applied to the present invention, and the preferred implementation methods and materials described in the specific implementation manners are for illustrative purposes only.

[0029] The following describes the specific method taking the monitoring of the movement behavior of Liza haematocheila in the tidal creeks of the Dongtan Wetland in the Yangtze Estuary as an example. The behavior monitoring methods of other estuarine fish such as Clupeidae, Mugil cephalus, Liza haematocheila, Engraulis japonicus, Sparidae, etc. are similar.

[0030] To further optimize and improve the case study on using unmanned aerial vehicles to monitor the migration behavior of Liza haematocheila in the Dongtan Wetland of the Yangtze Estuary, so as to more comprehensively demonstrate the technical details, data analysis methods and ecological significance of the research. The following is a more detailed description:

[0031] The Dongtan Wetland in the Yangtze Estuary is one of the typical estuarine wetlands in China and plays an important role in biodiversity protection, fish habitat provision and climate regulation. Tidal creeks are widely distributed in the Dongtan Wetland area, and these tidal creeks provide spawning, juvenile fish development and habitat sites for a variety of fish. The influence of tidal changes on the water depth and water quality of tidal creeks directly affects the activity range and behavior patterns of fish. As a typical fish species in estuarine wetlands, the migration and habitat behavior of Liza haematocheila are closely related to tidal changes. Studying the migration patterns and behavior habits of Liza haematocheila helps to protect the health of the estuarine wetland ecosystem and implement optimized protection measures.

[0032] Step 1: Define the monitoring area and applicable fish populations

[0033] The Dongtan Wetland in the Yangtze River Estuary is selected as the research area, specifically the tidal creek entrances and shallow water areas where fish schools often gather in the Yangtze River Estuary Wetland. Therefore, it is chosen as the starting point for UAV monitoring takeoff. The present invention selects the dominant species Liza haematocheila in the Yangtze River Estuary Wetland as the target fish school for monitoring. Liza haematocheila usually spawns in spring (April - May), prefers to gather and move in the shallow water areas of tidal creek entrances, and has a relatively long body length, which is visible to the naked eye. During high tide, Liza haematocheila usually enters deeper tidal creeks and coastal areas with the rising tide, and gradually returns to the vicinity of the tidal creek mouth during ebb tide, usually gathering in the area 2 to 8 kilometers away from the shore for short - distance migration activities. Therefore, it is representative as the target fish school for the Yangtze River Estuary Wetland.

[0034] Step 2: Preparation of UAV equipment and design of data collection

[0035] The present invention uses the DJI Mavic 2 Pro UAV for monitoring, equipped with a multi - spectral camera with a 20MP CMOS sensor, which can record 4K (3840×2160) high - definition video, and has a continuous flight time of about 30 minutes. The flight monitoring is selected during spring (the breeding period of Liza haematocheila) and during spring tides (April 24, 2024). The weather is clear and the wind speed is low during the aerial photography. The UAV initially starts from the coastline of the tidal creek mouth at a height of about 30 meters and flies about 200 meters to maximize the search area. When a school of Liza haematocheila is found at the tidal creek mouth, the flight height of the UAV is reduced and set at a height of 10 meters, and the flight speed is maintained at 1 - 3 m / s to ensure the clarity and breadth of the images and be able to record the behavior of Liza haematocheila in the shallow water area in detail. The orthogonal projection method is used during the flight process to minimize distortion and ensure the comparability of all videos.

[0036] Focus on observing the behavioral changes of Liza haematocheila at the lowest tide, the initial stage of rising tide, and the highest tide. Observations show that at the lowest tide, a large number of Liza haematocheila are observed to gather near the tidal creek mouth by the UAV, especially in the southern area of the tidal creek near the mudflat. As the rising tide begins, the tidal creek is gradually flooded by the tide, the water depth increases and the water quality is clear. Liza haematocheila tend to form groups of three to four and enter the tidal creek along the water flow from both sides of the tidal creek, moving inland along the water flow. Therefore, at the lowest tide, when Liza haematocheila gather at the tidal creek mouth, the UAV takes fixed - point and moving - tracking photos for 5 minutes. During the rising tide, when Liza haematocheila enter the tidal creek along the water flow, the UAV takes fixed - point and moving - tracking photos for 5 minutes, and each fixed - point and moving - tracking photo - taking duration lasts for 2 - 3 minutes respectively to ensure capturing key behaviors and covering all aspects of fish activities. In one tidal cycle, each behavior is photographed 3 - 5 times, and the activity behaviors of about 150 Liza haematocheila at the tidal creek mouth and in the tidal creek are recorded.

[0037] Step 3: Classification and analysis of fish behavior and processing of image data

[0038] The recorded videos were then imported into Kinovea (v.1.3) software. By comprehensively browsing all video content, the swimming behaviors of fish were visually screened and classified initially, and could be preliminarily classified into three types of behavioral patterns. According to this classification, three time intervals of video footage with the best fixed-point shooting were selected, each with a duration of about 1 minute, which could reflect each clear and complete fish movement behavior. Each video was clipped into segments of 20 to 40 seconds for detailed kinematic analysis.

[0039] Step 4: Reconstruction of the movement trajectories of fish populations and discrimination of behavior classification

[0040] Using the point tracking function of Kinovea, the individual fish in each segment in Step 3 were marked, and a continuous path was generated, which reflected the movement trajectories of the fish. The specific marking and tracking method was as follows: In the first frame, right-click on the position of the tracking object to be marked. After clicking, a scrolling menu will appear, and then select "Tracking Path". Then continue to right-click and select "Configure". A selection menu will appear. Click on the arrow next to "Measurement", then select "Speed" and click "Apply". If multiple fish need to be tracked, repeat the above steps. Finally, use the "Play" button to move the video segment forward, so that Kinovea can automatically track the selected object in the rest of the segment. After the playback is completed, right-click and select "Stop Tracking" to generate the movement trajectories of all marked fish ( Figure 1 ). After the tracking is completed, select "Linear Motion" in "Tools" in the top menu bar to analyze the displacement and speed of these trajectories ( Figure 1 ), and we can analyze and quantify the behavior of the fish. For example, medium-speed straight-line movement may indicate the "exploration" behavior of mullets, fast local turning-back movement may be related to the "resting" behavior of mullets, and slow straight-line followed by local turning-back movement may be related to the "foraging" behavior of mullets (Table 1):

[0041] Table 1 Quantification of mullet behavior and movement trajectories

[0042] Behavior Trajectory description Velocity curve Displacement curve Explore Show a straight-line motion trajectory The velocity fluctuates in the range of 0 - 240 px / s The vertical movement distance increases linearly Rest Relatively fixed in a local area, showing a small-range turning and circling trajectory The velocity fluctuates in the range of 0 - 320 px / s The vertical movement distance remains basically unchanged Forage Show a two-stage trajectory. First, it is a straight-line motion trajectory. After finding a suitable foraging area, it is relatively fixed in position and shows a small-range turning motion trajectory The velocity fluctuates in the range of 0 - 100 px / s The vertical movement distance increases linearly and then remains basically unchanged (after finding a suitable foraging area)

[0043] The parts not described in this invention are the same as the prior art or are implemented using the prior art. The applicant declares that this invention illustrates the detailed method of this invention through the above embodiments, but this invention is not limited to the above detailed method, that is, it does not mean that this invention must rely on the above detailed method to be implemented. Those skilled in the art should understand that any improvement to this invention, the equivalent replacement of each raw material of this invention's product, the addition of auxiliary components, and the selection of specific methods, etc., all fall within the protection scope and the disclosure scope of this invention.

Claims

1. A method for monitoring the movement behavior of fish near the estuary based on drones, characterized in that: The steps include: A. Clarify monitoring areas and applicable fish stocks The estuary and coastal areas were selected as the research and monitoring areas, with key areas including tidal creeks, shallows and areas with abundant aquatic plants; the dominant estuarine fish species that are suitable for drone observation at an altitude of 5-20 meters, often active on the water surface or in shallow waters near the coast, with clustered movements and visible body length, were selected as the research and monitoring areas; B. Data Collection Choose a drone with high-resolution imaging capabilities as a shooting tool; select a shooting time period based on the life cycle and tidal cycle of the target fish school, and shoot at high tide when the weather is clear based on the weather conditions; the shooting frequency is determined according to the activity habits of the fish and the tidal cycle, observe the activity trends of the fish school at the beginning of the tidal cycle, and arrange the corresponding tracking shooting frequency according to the different activity groups. When shooting, lower the flight altitude after searching for the target fish school and maintain a certain flying speed to ensure that all parts of the target fish's body are clearly visible in the field of vision; during a shooting process, it is necessary to arrange both fixed-point shooting and tracking mobile shooting, and the length of each shooting should ensure that key behaviors are captured and all aspects of fish activities are covered; C. Fish behavior classification analysis and image data processing Import the recorded video into a video analysis tool with point tracking function, and conduct a visual preliminary screening and classification of fish swimming behaviors by comprehensively browsing all video contents, that is, observe the sudden changes in fish reactions or behaviors and classify them into one category; based on this classification, select the appropriate time period of video material and edit it; D. Reconstruction of fish population movement trajectories and classification of behavior The point tracking function is used to mark the position of fish in the video to generate a continuous path that can reflect the movement trajectory of fish. By analyzing the time, distance and speed of these trajectories, the behavior categories of fish are analyzed and quantified.

2. The method for monitoring the movement behavior of fish in estuaries and nearshore areas based on drones according to claim 1 is characterized in that: In step B, the ideal time to monitor estuarine fish is spring and autumn, especially from March to mid-May and September to November, when fish activity is high and fish come to the estuary to feed or breed.

3. The method for monitoring the movement behavior of fish in estuaries and nearshore areas based on drones according to claim 1 is characterized in that: In step B, when filming, the drone initially departs from the coastline at an altitude of about 30 meters and flies about 200 meters to maximize the search area; after observing the school of fish, the drone is lowered to an altitude of less than 15 meters and the flight speed is maintained at 1-3m / s to observe the fish activities in more detail and ensure that all parts of the animal's body are clearly visible in the field of view; The camera was positioned orthogonally above the water surface to minimize distortion and ensure that all videos were comparable.

4. The method for monitoring the movement behavior of fish in estuaries and nearshore areas based on drones according to claim 1 is characterized in that: In step B, when shooting, fixed-point shooting is performed every 5-10 minutes to fully record each active group. Fixed-point shooting and tracking mobile shooting need to be arranged simultaneously during one shooting process, and each shooting duration should be maintained at 2-3 minutes to ensure that key behaviors are captured and all aspects of fish activities are covered; each movement behavior of the fish is repeatedly shot multiple times in a tidal cycle, focusing on observing the behavioral changes of the fish at the lowest tide, the early stage of high tide and the highest tide.

5. The method for monitoring the movement behavior of fish in estuaries and nearshore areas based on drones according to claim 1 is characterized in that: In step C, fish reactions or behaviors are classified into three categories: exploration, resting, and foraging through visual preliminary screening and classification; based on this classification, appropriate time periods of video material are selected, requiring that the video material period can reflect each clear and complete fish movement behavior, and the video material is edited into multiple 20 to 60 second segments for detailed kinematic analysis.

6. The method for monitoring the movement behavior of fish in estuaries and nearshore areas based on drones according to claim 1 is characterized in that: In step D, during the process of marking and tracking the selected fish or fish school, each frame of motion trajectory is corrected with the actual motion trajectory at any time to ensure the accuracy of the data.

7. The method for monitoring the movement behavior of fish in estuaries and nearshore areas based on unmanned aerial vehicles according to claim 1 is characterized in that: In step D, use Kinovea's point tracking function to generate a continuous path that reflects the fish's motion trajectory as follows: in the first frame, right-click to mark the location of the object you want to track, click and select "Tracking Path", then right-click again, select "Configure", click the arrow next to "Measurement", then select "Speed" and click "Apply"; repeat the above steps if you need to track multiple fish; finally, use the "Play" button to move the video clip forward so that Kinovea can automatically track the selected object in the rest of the clip; after playback is complete, right-click and select "Stop Tracking" to generate the motion trajectory of all marked fish.

8. The method for monitoring the movement behavior of fish in estuaries and nearshore areas based on drones according to claim 1 is characterized in that: In step D, the method for analyzing and quantifying the behavioral categories of fish is as follows: when the fish moves in a medium-speed straight line and the vertical movement distance increases linearly, it indicates the exploration form of the fish; when the fish moves rapidly and partially in a reciprocating manner and the vertical movement distance remains basically unchanged, it indicates the resting behavior of the fish; when the fish moves slowly and linearly and then partially in a reciprocating manner, the vertical movement distance increases linearly and then remains basically unchanged, it indicates the foraging behavior of the fish.

9. The method for monitoring the movement behavior of fish in estuaries and nearshore areas based on unmanned aerial vehicles according to claim 1 is characterized in that: In step D, the medium speed fluctuates in the range of 0-240px / s, the fast speed fluctuates in the range of 0-320px / s, and the slow speed fluctuates in the range of 0-100px / s.

Images