Fish counting method in industrial aquaculture scene

Abstract

The invention provides a fish counting method in an industrial aquaculture scene. The method comprises the following steps: carrying out conversion processing on input data; according to the diameter of the factory culture pond, sound ray data corresponding to the fixed pond wall needing to be removed are calculated; performing distance compensation on the echo sound ray data without the pool wall according to a distance compensation formula; environmental noise is removed according to fish echo characteristics; taking the average value of the current sound ray data as a threshold value, and performing vertex search on the part exceeding the threshold value; according to the average length and width of the fingerlings, the repeatedly counted fingerlings are removed; and according to the scanning range of the upper computer, performing fish quantity evaluation on the data of each frame. According to the invention, the method achieves the precise counting of fishes in a fixed factory scene through the analysis of the factory culture scene, reduces the labor cost, improves the statistical efficiency, and meets the demands of the continuously increasing underwater biomass statistics of the factory culture scene.

Description

Technical Field

[0001] This invention relates to the field of aquaculture technology, specifically to a method for counting fish in a factory farming setting. Background Technology

[0002] With the acceleration of industrial modernization, the demand for technologies such as real-time monitoring of underwater ecological environment, accurate assessment of aquaculture biomass, and intelligent early warning of aquaculture facility safety is constantly increasing, which directly promotes the rapid development of the high-end equipment market such as specialized fishing sonar, underwater robots, and Internet of Things monitoring systems.

[0003] Traditional factory farming methods rely primarily on manual counting of fish in ponds, which is time-consuming and labor-intensive. Furthermore, some fish species are highly sensitive to lighting from cameras. Acoustic technology, on the other hand, can effectively avoid damaging the fish, and its signal processing speed is much faster than image annotation. Therefore, there is an urgent need for a sound-based, precise fish counting technology suitable for factory farming scenarios. Summary of the Invention

[0004] To address the shortcomings of existing technologies, the present invention aims to provide a fish counting method for factory farming scenarios. By analyzing factory farming scenarios, it achieves accurate fish counting in fixed factory farming environments, reduces labor costs, and improves statistical efficiency, thereby meeting the growing demand for underwater biomass statistics in factory farming scenarios.

[0005] The technical solution of the present invention is as follows:

[0006] A method for counting fish in a factory farming setting includes the following steps:

[0007] Transform the input data;

[0008] Based on the diameter of the factory-style aquaculture pond, calculate the acoustic data corresponding to the fixed pond wall that needs to be removed;

[0009] Distance compensation is performed on echo ray data without pool walls according to the distance compensation formula;

[0010] Based on the echo characteristics of fish, environmental noise is removed;

[0011] Use the average value of the current vocal data as a threshold, and perform vertex lookup on the parts that exceed the threshold.

[0012] The number of duplicates was removed based on the average length and width of the fish species;

[0013] Based on the scanning range of the host computer, the number of fish in each frame of data is assessed.

[0014] Preferably, the step of converting the input data specifically includes: converting the bin file format input from the FPGA into decimal data that can be recognized by subsequent processing software.

[0015] Preferably, in the step of calculating the acoustic data corresponding to the fixed pool wall to be removed based on the diameter of the factory-style aquaculture pool, the pool wall and interference data exceeding the boundary are zeroed out.

[0016] Preferably, in the step of performing distance compensation on the echo ray data without a pool wall according to the distance compensation formula, the distance compensation formula is: , where a is the absorption coefficient and s is the distance relative to the sonar.

[0017] Preferably, in the step of using the average value of the current vocal data as a threshold and performing vertex lookup for the portion exceeding the threshold, the average value of the current vocal data is used as the threshold, and vertex lookup is performed for the portion exceeding the threshold. The found vertex is then moved down 3dB to find the endpoint, and a type determination is made based on a fuzzy function. When the number of data points within the endpoint is no more than 40 and greater than 25, it is considered a single target; if it exceeds 40, it is considered a group target. The determination value can be changed according to different fish species.

[0018] Preferably, the step of eliminating duplicate counts based on the average length and width of the fish species specifically includes:

[0019] The number of individual fish (Lsingle) is calculated based on a coefficient.

[0020] The group size Lgroup is calculated based on the coefficient.

[0021] The fish population assessment number for each vocal tract is: Ln = Lsingle + Lgroup.

[0022] Preferably, the step of evaluating the number of fish in each frame of data based on the scanning range of the host computer specifically includes: when the interface is scanning 360°, there are a total of 200 acoustic rays, then the output data is L. tot = L1+L2+L3+....+L 200 When the interface is scanned at 180°, there are a total of 100 sound lines, and the output data is L. tot = L1+L2+L3+....+L 100 .

[0023] Compared with existing technologies, the method of this invention analyzes the factory farming scenario, effectively removes interference factors such as pond walls, avoids the impact of environmental factors on statistical results, and ultimately achieves accurate fish counting in a fixed factory farming scenario. This reduces labor costs and improves statistical efficiency, in order to meet the growing demand for underwater biomass statistics in factory farming scenarios. Attached Figure Description

[0024] Other features, objects, and advantages of the present invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings:

[0025] Figure 1 This is a flowchart of the fish counting method in the factory farming scenario of the present invention;

[0026] Figure 2 This is a detailed flowchart of the fish counting method in the factory farming scenario of the present invention;

[0027] Figure 3 This is a schematic diagram of the pool wall calculation;

[0028] Figure 4 A schematic diagram comparing the target echo and noise signals;

[0029] Figure 5 A diagram illustrating vertex lookup;

[0030] Figure 6 This is a diagram illustrating endpoint lookup. Detailed Implementation

[0031] The present invention will now be described in detail with reference to specific embodiments. These embodiments will help those skilled in the art to further understand the present invention, but do not limit the invention in any way. It should be noted that those skilled in the art can make several changes and improvements without departing from the concept of the present invention. These all fall within the protection scope of the present invention.

[0032] Specifically, this invention provides a method for counting fish in a factory farming setting, such as... Figure 1 and Figure 2 As shown, the method includes the following steps:

[0033] S1: Perform transformation processing on the input data;

[0034] Specifically, the bin file format input to the FPGA is converted into decimal data that can be recognized by the subsequent processing software, and then calculations are performed on each vocal line.

[0035] S2: Calculate the acoustic data corresponding to the fixed pool wall that needs to be removed based on the diameter of the factory-style aquaculture pond;

[0036] Specifically, based on the diameter of the fixed factory-style aquaculture pond, the acoustic data corresponding to the fixed pond wall that needs to be removed is calculated, and the pond wall and interference data exceeding the boundary are zeroed out. Figure 3 As shown, in fast scan mode, the angle between two adjacent sound lines and the center B of the display control circle is 1.8°. BC is the 800 points in dB mode for one sound line. A is a point on the pool wall. α is the angle between the current sound line and the tangent, which can be calculated in conjunction with 1.8°. OB is the radius of the aquaculture pool. D is the diameter. Distance is the current display control range.

[0037] The specific calculation formula is as follows:

[0038] When the current pool wall's acoustic ray number is less than the acoustic ray number of 'l', the current pool wall's position in the data is:

[0039]

[0040] When the current pool wall's acoustic ray number is greater than the acoustic ray number of 'l', the current pool wall's position in the data is:

[0041]

[0042] When the current acoustic ray number of the pool wall is greater than the acoustic ray number of the estimated tangent, the position of the current pool wall in the data is:

[0043]

[0044] S3: Perform distance compensation on the echo ray data without pool walls according to the distance compensation formula;

[0045] The distance compensation formula is used to perform distance compensation on the echo ray data without pool walls. The distance compensation formula is as follows:

[0046]

[0047] Where a is the absorption coefficient, which is 200 dB / km, and s is the distance relative to the sonar.

[0048] S4: Remove environmental noise based on the echo characteristics of fish;

[0049] Because the acoustic echo and characteristics of the target differ from those of noise echo, environmental noise can be removed based on echo characteristics, such as... Figure 4 As shown.

[0050] S5: Use the average value of the current vocal data as a threshold, and perform vertex search on the part that exceeds the threshold;

[0051] Specifically, because the acquired echo data is affected by emission interference from the pool wall, it is necessary to use the average value of the current acoustic data as a threshold, and then search for vertices in the portion exceeding the threshold, such as... Figure 5 and Figure 6 As shown, the found vertex is moved down 3dB to find the endpoint, and the type is determined according to the fuzzy function. When the number of data in the endpoint does not exceed 40 and is greater than 25, it is a single target; if it exceeds 40, it is a group target. The judgment value can be changed according to different fish species.

[0052] S6: Based on the average length and width of the fish species, the number of duplicate counts is eliminated;

[0053] 1. Calculate the number of individual fish using a coefficient.

[0054] The estimated average body length of the fish is Lfish. In fast scan mode, the interval between two adjacent sound rays is α = 1.8°. Based on the radian formula and the current distance s of the target, the arc length L of the two sound rays at different distances from the sonar can be calculated. Then, based on the relationship between L and Lfish, the number of sound rays occupied by the target can be calculated.

[0055] When L>=Lfish, the count of the single entity being judged is 1;

[0056] When L < Lfish, the number of individual cells to be counted is L / Lfish.

[0057] 2. Calculate the group size Lgroup by coefficient.

[0058] The estimated average width of the fish is Lfish_vert, and the vertical distance between two adjacent fish is L0. The sonar vertical opening angle is approximately 12°. Therefore, the number of fish can be estimated based on the Pythagorean theorem and the distance s of the target group.

[0059] 2*(s*tan(β) / (Lfish_vert+L0))

[0060] The fish population assessment number for each vocal tract is: Ln = Lsingle + Lgroup.

[0061] S7: Based on the scanning range of the host computer, assess the number of fish in each frame of data.

[0062] Specifically, when the interface is scanning 360°, there are a total of 200 sound lines, and the output data is L. tot = L1+L2+L3+....+L 200 When the interface is scanned at 180°, there are a total of 100 sound lines, and the output data is L. tot = L1+L2+L3+....+L 100 .

[0063] Specific embodiments of the present invention have been described above. It should be understood that the present invention is not limited to the specific embodiments described above, and those skilled in the art can make various changes or modifications within the scope of the claims, which do not affect the essence of the present invention. Unless otherwise specified, the embodiments and features described in this application can be arbitrarily combined with each other.

Claims

1. A method for counting fish in a factory farming setting, characterized in that, The method includes the following steps: Transform the input data; Based on the diameter of the factory-style aquaculture pond, calculate the acoustic data corresponding to the fixed pond wall that needs to be removed; Distance compensation is performed on echo ray data without pool walls according to the distance compensation formula; Based on the echo characteristics of fish, environmental noise is removed; Use the average value of the current vocal data as a threshold, and perform vertex lookup on the parts that exceed the threshold. The number of duplicates was removed based on the average length and width of the fish species; Based on the scanning range of the host computer, the number of fish in each frame of data is assessed.

2. The fish counting method in a factory farming setting according to claim 1, characterized in that, The steps for converting the input data specifically include: converting the bin file format input from the FPGA into decimal data that can be recognized by subsequent processing software.

3. The fish counting method in a factory farming setting according to claim 1, characterized in that, In the step of calculating the acoustic data corresponding to the fixed pool wall that needs to be removed based on the diameter of the factory-style aquaculture pool, the pool wall and interference data beyond the boundary are set to zero.

4. The fish counting method in a factory farming setting according to claim 1, characterized in that, In the step of performing distance compensation on the echo ray data without a pool wall according to the distance compensation formula, the distance compensation formula is: , where a is the absorption coefficient and s is the distance relative to the sonar.

5. The fish counting method in a factory farming setting according to claim 1, characterized in that, In the step of using the average value of the current vocal data as a threshold and performing vertex search on the part exceeding the threshold, the average value of the current vocal data is used as a threshold, the part exceeding the threshold is searched for vertices, the found vertices are moved down 3dB to find the endpoints, and the type is determined according to the fuzzy function. When the number of data points within an endpoint is no more than 40 and greater than 25, it is considered a single target; if it exceeds 40, it is considered a group target. The judgment value can be changed according to different fish species.

6. The fish counting method in a factory farming setting according to claim 1, characterized in that, The step of eliminating duplicate counts based on the average length and width of the fish species specifically includes: The number of individual fish (Lsingle) is calculated based on a coefficient. The group size Lgroup is calculated based on the coefficient. The fish population assessment number for each vocal tract is: Ln = Lsingle + Lgroup.

7. The fish counting method in a factory farming setting according to claim 1, characterized in that, The step of evaluating the number of fish in each frame of data based on the scanning range of the host computer specifically includes: when the interface is scanning 360°, there are a total of 200 acoustic rays, and the output data is L. tot = L1+L2+L3+....+L 200 When the interface is scanned at 180°, there are a total of 100 sound lines, and the output data is L. tot = L1+L2+L3+....+L 100 .

Images