Water ecology intelligent monitoring device and method based on fish AI online identification

Abstract

The invention provides a water ecology intelligent monitoring device and method based on fish AI online identification, and relates to the technical field of video monitoring processing, the monitoring device comprises an angle regulation and control mechanism, an image acquisition unit, a cleaning mechanism and an induction detection assembly, the bottom of the angle regulation and control mechanism is provided with a top plate, and the bottom of the top plate is provided with a floating plate in an attached mode; the bottom end of the angle regulation and control mechanism is connected with a rotating shaft, the outer side of the rotating shaft is sleeved with an inner-layer sleeve, the outer side of the inner-layer sleeve is sleeved with an outer-layer sleeve, and the rotating shaft and the outer-layer sleeve rotate. According to the invention, the image acquisition unit at the bottom can be cleaned flexibly, a generated image acquisition dead angle area is eliminated, the image acquisition unit can be sealed and protected when identification and acquisition are not carried out for a long time, and meanwhile, whether fishes approach the surrounding environment or not can be detected comprehensively.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of video monitoring processing, in particular to a water ecological intelligent monitoring device and method based on fish AI online identification. BACKGROUND

[0002] The water ecological monitoring technology based on fish AI identification collects fish activity and water quality data in real time through an underwater camera array and an Internet of Things sensor, and performs localized AI processing in combination with an edge computing terminal. Parameters such as dissolved oxygen and turbidity can be analyzed synchronously, a fish biological integrity index (F-IBI) is constructed to evaluate ecological health, and the monitoring reliability in turbid water and at night is improved through multi-modal data fusion technology. The technology supports illegal fishing identification, invasive species early warning, and other scenarios, and provides dynamic decision support for water ecological protection.

[0003] In the prior art, the water ecological intelligent monitoring scheme based on fish AI requires multiple image collection and monitoring devices to be built underwater to obtain raw data of the surrounding fish ecology. However, due to the diverse types of underwater environments, the monitoring device will accumulate a large amount of impurities on the surface when it is submerged for a long time, thereby reducing the clarity of the obtained images. Directly building a cleaning system on the side will cause a certain obstruction to the image collection range, resulting in a dead angle area for identification. On the other hand, conventional monitoring devices cannot fully probe whether fish are approaching from all around, and installing multiple vibration sensors will increase the cost of the device and the overall failure rate of the device when installed underwater, requiring frequent maintenance. SUMMARY

[0004] To address the deficiencies of the prior art, the present application aims to provide a water ecological intelligent monitoring device and method based on fish AI online identification to solve the problems raised in the background art. The present application can construct a fish biological integrity index to comprehensively evaluate ecological health, achieve rapid response to pollution events, flexibly provide cleaning treatment for the image collection unit at the bottom, eliminate the generated image collection dead angle area, protect the image collection unit when not performing identification collection for a long time, and fully probe whether fish are approaching from all around.

[0005] In order to achieve the above object, the present application is realized by the following technical scheme: A water ecological intelligent monitoring device based on fish AI online identification, comprising a monitoring device body, the monitoring device body comprises an angle control mechanism, an image acquisition unit, a cleaning mechanism and a sensing detection assembly, the bottom of the angle control mechanism is provided with a top plate, the bottom of the top plate is attached with a floating plate, the bottom end of the angle control mechanism is connected with a rotating shaft, the outer side of the rotating shaft is sleeved with an inner sleeve, the outer side of the inner sleeve is sleeved with an outer sleeve, the rotating shaft and the outer sleeve are in rotary motion, the inner sleeve always remains in a fixed state, the bottom end of the rotating shaft is embedded into the inside of the image acquisition unit, the bottom of the outer sleeve is connected to the top end of the cleaning mechanism, the bottom of the cleaning mechanism is integrally formed with a bottom plate, the surface of the bottom plate is embedded with a plurality of sensing detection assemblies, and the sensing detection assemblies are evenly distributed above the bottom plate, and the bottom of the floating plate is embedded with a plurality of floating columns.

[0006] Further, the angle control mechanism comprises a driving bin, a first gear and a second gear, the inside of the driving bin is screwed with a motor, the surface of the top plate is inserted with a fixed seat, the top of the fixed seat is integrally formed with a fixed plate, and the top of the rotating shaft is embedded into the bottom of the fixed plate through a bearing.

[0007] Further, the surface of the rotating shaft is key connected with the second gear, the second gear is pressed on the surface of the top plate, the outer sleeve is welded on the bottom surface of the second gear, and the outer sleeve passes from the middle position of the floating plate and the top plate upwards, the side of the first gear is sleeved with a toothed belt, and the toothed belt and the second gear are engaged with the transmission gear inside the driving bin.

[0008] Further, the top of the second gear is provided with a butt joint sleeve ring, the bottom of the fixed plate is welded with a plug-in sleeve, the bottom of the plug-in sleeve is embedded into the inside of the butt joint sleeve ring, and the bottom of the plug-in sleeve is pressed on the bottom of the butt joint sleeve ring through a steel ball.

[0009] Further, the image acquisition unit comprises an image acquisition camera and a sealed bin, the image acquisition camera is screwed at the end of the rotating shaft, the sealed bin is welded at the bottom of the inner sleeve, the side of the sealed bin is provided with a light transmission plate, a limit ring is attached to the inner wall of the sealed bin, a threaded sleeve is formed at the bottom of the sealed bin, and the image acquisition camera is embedded into the inside of the sealed bin.

[0010] Further, the cleaning mechanism comprises a transmission arm, a sponge sleeve and a rotating bin, the middle of the transmission arm is integrally formed at the bottom of the outer sleeve, the end of the transmission arm is inserted with an expansion pipeline, the sponge sleeve is sleeved on the surface of the expansion pipeline, and the bottom end of the expansion pipeline is welded with a rotating bin.

[0011] Further, the top of the rotating bin is provided with a rubber scraping strip, the inner side of the rubber scraping strip and the inner wall of the rotating bin are aligned with the outer surface of the light-transmitting plate, a threaded column is inserted into the inner side of the rotating bin, the threaded column is inserted into the inner side of the threaded sleeve, and the side of the sponge sleeve is pressed on the surface of the light-transmitting plate.

[0012] Further, the inductive detection assembly comprises a base, a conductive ball and a floating ball, a clamping groove is formed in the top of the base, and a stand is integrally formed on the top of the base on both sides, and a conductive ring is installed on the top of the stand.

[0013] Further, the conductive ball is embedded in the inner side of the clamping groove, the top of the conductive ball is connected with a conductive wire, the top of the conductive wire is connected with a floating ball, and the conductive ball rotates in the inner side of the clamping groove.

[0014] A monitoring method using the above monitoring device, comprising the following steps:

[0015] Step one, build an underwater image acquisition array;

[0016] Step two, determine the fish direction through the inductive detection unit to obtain comprehensive fish image information;

[0017] Step three, the monitoring device is integrated with water quality sensors such as dissolved oxygen, pH, turbidity, conductivity and ammonia nitrogen at the bottom, and data is collected once in a fixed period;

[0018] Step four, automatically track the fish school through the moving target detection algorithm, calculate the population density and biomass, and identify abnormal behavior;

[0019] Step five, ecological health assessment, construct fish biological integrity index model, combine species richness and specific species ratio to output ecological health rating;

[0020] Step six, regularly clean the monitoring device, and establish a remote diagnosis system to predict equipment failure through a vibration sensor.

[0021] The beneficial effects of the present application are:

[0022] 1. The water ecological intelligent monitoring device based on fish AI online identification can flexibly provide cleaning treatment for the image acquisition unit at the bottom, eliminate the generated image acquisition dead angle area, and protect the image acquisition unit when not identifying and collecting for a long time, thereby further improving the clarity of the image acquisition in the use state.

[0023] 2. The water ecological intelligent monitoring device based on fish AI online identification can comprehensively investigate whether there are fish approaching around the environment, expand the range of collected image data, make the monitoring angle more comprehensive, avoid greatly increasing the equipment cost, and have a lower failure rate.

[0024] 3. The fish AI online identification-based water ecological intelligent monitoring method fuses underwater images, sonar and water quality parameters (dissolved oxygen / turbidity, etc.), constructs a fish biological integrity index (F-IBI) to comprehensively evaluate ecological health, realizes rapid response to pollution events through multi-modal data linkage (such as fish abnormal behavior + water quality mutation), and breaks through the turbid water and night monitoring bottleneck through edge computing and cross-modal algorithm, so that the accuracy is effectively improved compared with traditional optical means. BRIEF DESCRIPTION OF DRAWINGS

[0025] Figure 1 It is a structural schematic view of the appearance of the fish AI online identification-based water ecological intelligent monitoring device of the application;

[0026] Figure 2 It is a structural schematic view of the transmission part of the angle regulating mechanism to the cleaning mechanism;

[0027] Figure 3 It is a structural schematic view of the transmission part of the angle regulating mechanism to the image acquisition camera;

[0028] Figure 4 It is a connection diagram of the image acquisition unit and the cleaning mechanism part;

[0029] Figure 5 It is an internal structure diagram of the image acquisition unit;

[0030] Figure 6 It is a schematic view of the cleaning mechanism;

[0031] Figure 7 It is a structural schematic view of the sensing detection assembly part;

[0032] Figure 8 It is a flowchart of the fish AI online identification-based water ecological intelligent monitoring method of the application;

[0033] In the figure: 1, floating plate; 2, angle regulating mechanism; 3, image acquisition unit; 4, cleaning mechanism; 5, bottom plate; 6, sensing detection assembly; 7, top plate; 8, floating column; 9, driving bin; 10, first gear; 11, outer sleeve; 12, butt joint collar; 13, fixed plate; 14, fixed seat; 15, rotating shaft; 16, plug-in sleeve; 17, second gear; 18, toothed belt; 19, image acquisition camera; 20, inner sleeve; 21, transmission arm; 22, sealing bin; 23, light transmission plate; 24, threaded sleeve; 25, limiting ring; 26, telescopic pipeline; 27, sponge sleeve; 28, rotating bin; 29, rubber scraping strip; 30, threaded column; 31, base; 32, clamping groove; 33, stand column; 34, conductive ring; 35, conductive ball; 36, conductive wire; 37, floating ball. DETAILED DESCRIPTION

[0034] In order to make the technical means, creative features, purposes and effects of the present application easy to understand, the present application is further described below in conjunction with specific embodiments.

[0035] Please refer to Figures 1 to 8 The present application provides the following technical solutions: an aquatic ecological intelligent monitoring device based on fish AI online identification, comprising a monitoring device body, the monitoring device body comprising an angle control mechanism 2, an image acquisition unit 3, a cleaning mechanism 4 and a sensing detection assembly 6, the bottom of the angle control mechanism 2 is provided with a top plate 7, the bottom of the top plate 7 is attached with a floating plate 1, the bottom end of the angle control mechanism 2 is connected with a rotating shaft 15, the outer side of the rotating shaft 15 is sleeved with an inner sleeve 20, the outer side of the inner sleeve 20 is sleeved with an outer sleeve 11, the rotating shaft 15 and the outer sleeve 11 are in rotary motion, the inner sleeve 20 always remains in a fixed state, the bottom end of the rotating shaft 15 is embedded into the inside of the image acquisition unit 3, the bottom of the outer sleeve 11 is connected to the top end of the cleaning mechanism 4, the bottom of the cleaning mechanism 4 is integrally formed with a bottom plate 5, the surface of the bottom plate 5 is embedded with a plurality of sensing detection assemblies 6, and the sensing detection assemblies 6 are evenly distributed above the bottom plate 5, and the bottom of the floating plate 1 is embedded with a plurality of floating columns 8. The monitoring device collects and identifies underwater images in the form of monitoring.

[0036] When the present application is used, underwater image acquisition arrays are built, each acquisition array is composed of a plurality of monitoring devices, each monitoring device is floated on the water surface by the floating plate 1 at the top, and the image acquisition unit 3 and the cleaning mechanism 4 are placed into the underwater area; a plurality of sensing detection assemblies 6 are arranged on the bottom of each monitoring device through the bottom plate 5, the fish direction is judged through the sensing detection unit, and the top angle control mechanism 2 is controlled to operate through the detected fish approaching information, and then the image acquisition unit 3 is rotated by the angle control mechanism 2, until the fish approaching is recognized by the fish AI, so as to collect comprehensive fish image information. The outer side of the image acquisition unit 3 is cleaned regularly by the cleaning mechanism 4, so as to ensure that the definition of the image data is sufficient, and when long-time monitoring is not required, the side of the image acquisition unit 3 is also enclosed by the cleaning mechanism 4.

[0037] The angle control mechanism 2 comprises a driving bin 9, a first gear 10 and a second gear 17. The inside of the driving bin 9 is screwed with a motor. The surface of the top plate 7 is inserted with a fixed seat 14. The top of the fixed seat 14 is integrally formed with a fixed plate 13. The top of the rotating shaft 15 is embedded into the bottom of the fixed plate 13 through a bearing. The surface of the rotating shaft 15 is keyed with the second gear 17. The second gear 17 is pressed against the surface of the top plate 7. The outer sleeve 11 is welded at the bottom of the second gear 17. The outer sleeve 11 passes through the middle position of the floating plate 1 and the top plate 7 upwards. The side of the first gear 10 is sleeved with a toothed belt 18. The toothed belt 18 and the second gear 17 are engaged with the transmission gear inside the driving bin 9. The top of the second gear 17 is provided with a butt joint sleeve ring 12. The bottom of the fixed plate 13 is welded with a plug-in sleeve 16. The bottom of the plug-in sleeve 16 is embedded into the inside of the butt joint sleeve ring 12. The bottom of the plug-in sleeve 16 is pressed against the bottom of the butt joint sleeve ring 12 through a steel ball.

[0038] Specifically, after starting the motor inside the driving bin 9, the whole angle control mechanism 2 is driven to rotate by the motor bottom drive shaft. Through the multiple gear structures on the motor drive shaft, the rotation of the toothed belt 18 is controlled. Then the rotating shaft 15 and the bottom image acquisition camera 19 are driven to rotate through the linkage of the second gear 17, so as to change the image acquisition range. At the same time, the first gear 10 on the outside is driven to rotate through the gear sets above and below the drive shaft. The bottom outer sleeve 11 is driven to rotate through the rotation of the first gear 10. The outer sleeve 11 directly drives the lower cleaning mechanism 4 to rotate, so as to realize the cleaning or switching process of the image acquisition unit 3. Since the cleaning mechanism 4 and the image acquisition camera 19 part always rotate synchronously, but the rotation speed and direction of the two are different, when the image acquisition camera 19 is shielded by the sponge sleeve 27 on the side of the cleaning mechanism 4 after rotating, the image acquisition camera 19 only needs to be continuously controlled to rotate. After rotating again, the image acquisition camera 19 returns to the same acquisition angle position. The sponge sleeve 27 has been rotated to other positions due to the different rotation speeds, so as to avoid the shielding problem.

[0039] The image acquisition unit 3 comprises an image acquisition camera 19 and a sealed bin 22. The image acquisition camera 19 is screwed at the end of the rotating shaft 15. The sealed bin 22 is welded at the bottom of the inner sleeve 20. The side of the sealed bin 22 is provided with a light-transmitting plate 23. The inner wall of the sealed bin 22 is provided with a limiting ring 25. The bottom of the sealed bin 22 is provided with a threaded sleeve 24. The image acquisition camera 19 is embedded in the sealed bin 22. The cleaning mechanism 4 comprises a transmission arm 21, a sponge sleeve 27 and a rotating bin 28. The middle of the transmission arm 21 is integrally formed at the bottom of the outer sleeve 11. The end of the transmission arm 21 is provided with an extension pipeline 26. The sponge sleeve 27 is sleeved on the surface of the extension pipeline 26. The bottom end of the extension pipeline 26 is welded with the rotating bin 28. The bottom image acquisition unit 3 can be flexibly cleaned. The generated image acquisition dead angle area is eliminated. The image acquisition unit 3 can be closed and protected for a long time without identification acquisition. The image acquisition clarity in the use state is further improved.

[0040] Specifically, after the outer sleeve 11 is rotated by the angle control mechanism 2, the transmission arm 21 at the bottom of the outer sleeve 11 drives the extension pipeline 26, the sponge sleeve 27 and the rotating bin 28 at the bottom to rotate. The sponge sleeve 27 scrapes the light-transmitting plate 23 at the side of the image acquisition unit 3 to achieve cleaning. When the image acquisition unit 3 needs to be completely closed, the entire cleaning mechanism 4 is controlled to continuously rotate by the angle control mechanism 2. The threaded column 30 at the bottom is embedded in the threaded sleeve 24. The rotating bin 28 moves up. The extension pipeline 26 shrinks. The rubber scraper 29 moves up along the surface of the light-transmitting plate 23. Finally, the rotating bin 28 is sleeved on the side of the sealed bin 22. The rubber layer on the inner wall of the rotating bin 28 completely extrudes and adheres to the surface of the light-transmitting plate 23. The internal image acquisition unit 3 is surrounded and closed.

[0041] The top of the rotating bin 28 is provided with a rubber scraping strip 29, the inner side of the rubber scraping strip 29 and the inner wall of the rotating bin 28 are aligned with the outer surface of the light transmission plate 23, the inner side of the rotating bin 28 is inserted with a threaded column 30, the top of the threaded column 30 is inserted into the inside of the threaded sleeve 24, and the side edge of the sponge sleeve 27 is pressed on the surface of the light transmission plate 23. The induction detection assembly 6 comprises a base 31, a conductive ball 35 and a floating ball 37, the top of the base 31 is provided with a clamping groove 32, the top end of the base 31 is integrally formed with a stand 33 on both sides, and the top end of the stand 33 is provided with a conductive ring 34. The clamping groove 32 is embedded with the conductive ball 35, the top of the conductive ball 35 is connected with a conductive wire 36, the top of the conductive wire 36 is connected with the floating ball 37, and the conductive ball 35 rotates in the clamping groove 32. Whether the surrounding environment has fish approaching can be comprehensively investigated, the image data range collected is expanded, the monitoring angle is more comprehensive, the equipment cost is not greatly increased, and the failure rate is lower.

[0042] Specifically, the bottom plate 5 distributes a plurality of induction detection assemblies 6 above, the floating ball 37 pulls the conductive wire 36 to always keep vertical upward under the condition that no external force is disturbed, when fish approach around, the generated corrugation affects the floating ball 37, causing the floating ball 37 to pull the conductive wire 36 at the bottom to rotate and deviate to one side, until the conductive wire 36 is in contact with the conductive ring 34 at the side edge, realizing the connection of the circuit, thereby generating an electric signal, and the image acquisition unit 3 is regulated and controlled by means of the electric signal, so as to rotate the image acquisition camera 19 to image acquisition of the approaching fish.

[0043] The embodiment also provides a monitoring method using the above monitoring device, comprising the following steps:

[0044] Step one, build an underwater image acquisition array, each acquisition array is composed of a plurality of monitoring devices, each monitoring device is floated on the water surface by the floating plate 1 at the top, and the image acquisition unit 3 and the cleaning mechanism 4 are placed into the underwater area;

[0045] Step two, a plurality of induction detection assemblies 6 are arranged on the bottom of each monitoring device through the bottom plate 5, the fish direction is judged through the induction detection unit, and the top angle regulation mechanism 2 is controlled to operate through the detected fish approaching information, and then the image acquisition unit 3 is rotated by means of the angle regulation mechanism 2, until the approaching fish is recognized by the fish AI, so as to collect comprehensive fish image information;

[0046] Step three, the monitoring device is integrated with water quality sensors such as dissolved oxygen, pH, turbidity, conductivity and ammonia nitrogen at the bottom, and data is collected once in a fixed period, and the monitoring device is deployed in non-flowing areas such as lakes and reservoirs;

[0047] Step four, automatically track the fish school through the moving target detection algorithm, calculate the population density and biomass, identify abnormal behavior, and establish a multi-parameter linkage response mechanism: when the dissolved oxygen is less than 4 mg / L and a large number of small fish die, automatically trigger the heavy metal pollution early warning, and locate the pollution source probability distribution area;

[0048] Step five, ecological health assessment, construct fish biological integrity index model, combine species richness and unique species ratio to output ecological health rating, input visible light image and side scan sonar data into multi-modal Transformer network to solve the identification problem in turbid water environment, and improve the accuracy of night monitoring;

[0049] Step six, real-time chain key monitoring data, generate an unalterable ecological audit report to meet the ESG information disclosure requirements. Regularly clean the monitoring device and establish a remote diagnosis system to predict equipment failure through vibration sensors.

[0050] This method combines underwater images, sonar and water quality parameters (dissolved oxygen / turbidity, etc.), constructs a fish biological integrity index (F-IBI) to comprehensively evaluate ecological health, and realizes rapid response to pollution events through multi-modal data linkage (such as fish school abnormal behavior + water quality mutation). Edge computing and cross-modal algorithm break through the bottleneck of turbid water and night monitoring, and the accuracy is effectively improved compared with traditional optical means.

[0051] The above shows and describes the basic principles and main features of the present application and the advantages of the present application. For those skilled in the art, it is obvious that the present application is not limited to the details of the above exemplary embodiments, and the present application can be realized in other specific forms without departing from the spirit or essential characteristics of the present application.

[0052] In addition, it should be understood that although the present specification is described in terms of embodiments, each embodiment does not contain only one independent technical solution, and the description manner of the specification is only for the sake of clarity. Those skilled in the art should consider the specification as a whole, and the technical solutions in each embodiment can be appropriately combined to form other embodiments that those skilled in the art can understand.

Claims

1. An intelligent water ecology monitoring device based on AI online fish identification, comprising a monitoring device body, characterized by: The monitoring device body comprises an angle control mechanism (2), an image acquisition unit (3), a cleaning mechanism (4) and an induction detection assembly (6); a top plate (7) is installed at the bottom of the angle control mechanism (2); a floating plate (1) is attached to the bottom of the top plate (7); a rotating shaft (15) is connected to the bottom end of the angle control mechanism (2); an inner sleeve (20) is sleeved on the outer side of the rotating shaft (15); an outer sleeve (11) is sleeved on the outer side of the inner sleeve (20); the rotating shaft (15) and the outer sleeve (11) are sleeved on the outer side of the rotating shaft (15); 1) both perform rotational motion, the inner sleeve (20) always remains in a fixed state, the bottom end of the rotating shaft (15) is embedded in the interior of the image acquisition unit (3), the bottom of the outer sleeve (11) is connected to the top of the cleaning mechanism (4), the bottom of the cleaning mechanism (4) is integrally formed with a bottom plate (5), the surface of the bottom plate (5) is embedded with a plurality of induction detection components (6), and the induction detection components (6) are evenly distributed above the bottom plate (5), and the bottom of the floating plate (1) is embedded with a plurality of floating columns (8).

2. The intelligent water ecology monitoring device based on AI online fish identification according to claim 1 is characterized by: The angle control mechanism (2) comprises a drive chamber (9), a first gear (10) and a second gear (17); a motor is screwed into the interior of the drive chamber (9); a fixing seat (14) is inserted into the surface of the top plate (7); a fixing plate (13) is integrally formed on the top of the fixing seat (14); and the top of the rotating shaft (15) is embedded in the bottom of the fixing plate (13) through a bearing.

3. The intelligent water ecology monitoring device based on AI online fish identification according to claim 2 is characterized by: The surface key of the rotating shaft (15) is connected to the second gear (17), and the second gear (17) is pressed on the surface of the top plate (7). The outer sleeve (11) is welded to the bottom surface of the second gear (17), and the outer sleeve (11) passes upward from the middle position of the floating plate (1) and the top plate (7). The side of the first gear (10) is provided with a toothed belt (18), and the toothed belt (18) and the second gear (17) are both engaged with the transmission gear inside the drive chamber (9).

4. The intelligent water ecology monitoring device based on AI online fish identification according to claim 3 is characterized by: A docking ring (12) is provided on the top of the second gear (17), and a plug-in sleeve (16) is welded to the bottom of the fixing plate (13). The bottom of the plug-in sleeve (16) is embedded in the inner side of the docking ring (12), and the bottom of the plug-in sleeve (16) is pressed against the bottom of the docking ring (12) by a steel ball.

5. The intelligent water ecology monitoring device based on AI online fish identification according to claim 2 is characterized by: The image acquisition unit (3) comprises an image acquisition camera (19) and a sealed chamber (22), wherein the image acquisition camera (19) is screwed to the end of the rotating shaft (15), the sealed chamber (22) is welded to the bottom of the inner sleeve (20), and a light-transmitting plate (23) is installed on the side of the sealed chamber (22), a limiting ring (25) is mounted on the inner wall of the sealed chamber (22), a threaded sleeve (24) is provided at the bottom of the sealed chamber (22), and the image acquisition camera (19) is embedded in the interior of the sealed chamber (22).

6. The intelligent water ecology monitoring device based on AI online fish identification according to claim 5 is characterized by: The cleaning mechanism (4) comprises a transmission arm (21), a sponge sleeve (27) and a rotating chamber (28); the middle of the transmission arm (21) is integrally formed at the bottom of the outer sleeve (11); a telescopic pipe (26) is inserted into the bottom end of the transmission arm (21); the sponge sleeve (27) is sleeved on the surface of the telescopic pipe (26); and the bottom end of the telescopic pipe (26) is welded with a rotating chamber (28).

7. The intelligent water ecology monitoring device based on AI online fish identification according to claim 6 is characterized by: A rubber scraper (29) is provided on the top of the rotating bin (28), and the inner side of the rubber scraper (29) and the inner wall of the rotating bin (28) are aligned with the outer surface of the light-transmitting plate (23). A threaded column (30) is inserted on the inner side of the rotating bin (28), and the top of the threaded column (30) is inserted into the inside of the threaded sleeve (24), and the side of the sponge sleeve (27) is pressed against the surface of the light-transmitting plate (23).

8. The intelligent water ecology monitoring device based on AI online fish identification according to claim 6 is characterized by: The inductive detection assembly (6) comprises a base (31), a conductive ball (35) and a floating ball (37); a slot (32) is provided on the top of the base (31); columns (33) are integrally formed on both sides of the top of the base (31); and a conductive ring (34) is installed on the top of the column (33).

9. The intelligent water ecology monitoring device based on AI online fish identification according to claim 8, characterized in that: A conductive ball (35) is embedded in the slot (32), a conductive wire (36) is connected to the top of the conductive ball (35), a floating ball (37) is connected to the top of the conductive wire (36), and the conductive ball (35) rotates inside the slot (32).

10. A monitoring method using the monitoring device according to claim 1, characterized in that: The following steps are involved: Step 1: Build an underwater image acquisition array; Step 2: Determine the fish's position through the sensing detection unit to obtain comprehensive fish image information; Step 3: The bottom of the monitoring device is integrated with water quality sensors such as dissolved oxygen, pH, turbidity, conductivity, and ammonia nitrogen, and data is collected once at a fixed period; Step 4: Automatically track fish schools using motion target detection algorithms, calculate population density and biomass, and identify abnormal behavior; Step 5: Conduct an ecological health assessment, construct a fish biomass integrity index model, and output an ecological health rating based on species richness and endemic species proportion; Step 6: Regularly perform self-cleaning on the monitoring device and establish a remote diagnosis system to predict equipment failures through vibration sensors.

Images