Sensing fishing ground prediction system and method

TWI945542BActive Publication Date: 2026-11-11NATIONAL TAIWAN OCEAN UNIVERSITY
View PDF 3 Cites 0 Cited by

Patent Information

Application Number
TW113124110
Authority / Receiving Office
TW · TW
Patent Type
Patents
Current Assignee / Owner
Filing Date
2024-06-27
Publication Date
2026-11-11
Estimated Expiration
2044-06-26

Smart Images

  • Figure TWG2TB001920287_001
    Figure TWG2TB001920287_001
  • Figure TWG2TB001920287_002
    Figure TWG2TB001920287_002
  • Figure TWG2TB001920287_003
    Figure TWG2TB001920287_003
Patent Text Reader

Abstract

A remote sensing fishing ground prediction system is disclosed, comprising a terminal device for transmitting at least one selected fishing ground information within a selected area to a user, and including a plurality of sensing data acquisition units, a storage unit, and a computing device. The sensing data acquisition units are located within the selected area and perform environmental sensing to generate a plurality of current sensing data at a plurality of locations. The storage units store the plurality of current sensing data and a plurality of historical sensing data at the plurality of locations. The computing device establishes a fishing ground prediction model for the selected area based on the plurality of historical sensing data, and uses the fishing ground prediction model to predict a probability model of fish school distribution in the selected area within a specific time period based on the plurality of current sensing data. Then, based on the fish school distribution probability model, the system generates selected fishing ground information and transmits it to the user's terminal device.
Need to check novelty before this filing date? Find Prior Art

Description

[Technical Field]

[0001] This invention relates to a remote sensing fishery prediction system, and more particularly to a remote sensing fishery prediction system that can predict the probability of fish distribution in a selected field by using a plurality of current sensing data from a plurality of locations in the selected field. [Previous Technology]

[0002] Fishing is a high-risk profession. Fishermen need to work at sea for long periods of time. Since fish are usually easier to catch at night or in rainy weather, fishermen often have to work at night or in rainy weather in order to catch more fish. Therefore, fishing boats are usually equipped with satellite positioning devices to increase the safety of the boat.

[0003] Regarding marine satellite positioning systems, they are generally used to locate the current or expected destination of vessels such as fishing boats, cargo ships, passenger ships, oil tankers, or warships. The main function of current marine positioning systems is to guide the direction of travel of ships, allowing them to navigate the vast ocean to the correct destination or avoid dangerous waters.

[0004] However, general marine satellite positioning systems cannot help fishermen catch more fish. If fishermen want to have a better catch, they still have to rely on their own experience to judge the weather, ocean currents and sea conditions. Moreover, a fishing trip may not always result in a return empty-handed. Although more and more fishermen are now using tools such as "fish finders" and "fish detectors" that use sound waves to detect the location of fish schools, these tools require the fishing boat to keep searching at sea to find the location of the fish schools. If the possible location of the fish schools can be predicted in advance, the detection time can be greatly reduced and the catch can be increased. [Summary of the Invention]

[0005] In view of this, in order to solve the problems encountered above, the inventor, based on years of practical experience and rich related knowledge, invented a remote sensing fishing ground prediction system. The remote sensing fishing ground prediction system can use the communication nodes of fishing boats, satellites, and drones to obtain environmental sensing data from environmental sensors set up at sea as sensing data acquisition units. The system then uses a computing processing center as a computing device to build a fishing ground prediction model based on multiple historical sensing data. The system then uses the fishing ground prediction model to predict the probability of fish distribution in a selected area based on multiple current sensing data. This prediction data is transmitted to the user's terminal device, allowing the user to obtain the predicted location of fish within a specific time period and to go to that location to catch fish within that specific time period, greatly improving work efficiency and the success rate of fishing.

[0006] Based on at least one objective of the present invention, the present invention provides a remote sensing fishery prediction system for transmitting at least one piece of selected fishery information within a selected field to a user's terminal device. The remote sensing fishery prediction system includes a plurality of sensing data acquisition units, a storage unit, and a computing device. The plurality of sensing data acquisition units are disposed at a plurality of locations within the selected field and are used to perform environmental sensing to generate a plurality of current sensing data at the plurality of locations. The storage unit is communicatively connected to the plurality of sensing data acquisition units and is used to store the plurality of current sensing data and a plurality of historical sensing data at the plurality of locations. The computing device is communicatively connected to the storage unit and the terminal device and is used to establish a fishery prediction model for the selected field based on the plurality of historical sensing data. Then, the plurality of current sensing data at the plurality of locations are input into the fishery prediction model to predict the probability model of fish school distribution in the selected field within a specific time period. Finally, the selected fishery information is generated and transmitted to the terminal device based on the probability model of fish school distribution. The specific time period refers to the time required for the evolution of the food chain in the selected field.

[0007] In one embodiment of the present invention, the computing device is used to run software programs to configure a data preprocessing module and a learning algorithm module, wherein the data preprocessing module is signal-connected to the learning algorithm module.

[0008] In one embodiment of the present invention, in prediction mode, the data preprocessing module is used to preprocess a plurality of current sensing data, and the learning algorithm module will bring the preprocessed plurality of current sensing data into the fishery prediction model to predict the probability model of fish distribution in a selected field within a specific time period.

[0009] In one embodiment of the present invention, in training mode, the data preprocessing module is used to preprocess a plurality of historical sensing data, and the learning algorithm module then establishes a fishing ground prediction model based on the plurality of historical sensing data.

[0010] In one embodiment of the present invention, the data preprocessing module uses one or a combination of genetic algorithm, data clustering, image processing, edge detection, chi-square test, EM algorithm (Expectation-Maximization Algorithm), and data labeling for data preprocessing.

[0011] In one embodiment of the present invention, the learning algorithm module determines multiple water color data for multiple locations based on multiple current sensing data and multiple historical sensing data for multiple locations, and the fishery prediction model predicts the probability model of fish distribution in a selected field within a specific time period based on multiple water color data for multiple locations.

[0012] In one embodiment of the present invention, the learning algorithm module has a water color data correction formula, which is used to correct a plurality of current sensing data and a plurality of historical sensing data at a plurality of locations, thereby correcting a plurality of water color data at a plurality of locations, and the fishery prediction model predicts the probability model of fish distribution in a selected field at a specific time based on the corrected plurality of water color data at a plurality of locations.

[0013] In one embodiment of the present invention, the remote sensing fishing ground prediction system further includes a plurality of communication node devices, which are communicatively connected to a plurality of sensing data acquisition units and storage units, wherein each communication node device is mounted on one of an artificial satellite, a fishing boat, a drone, and a vehicle.

[0014] In one embodiment of the present invention, the computing device further has a pricing formula, and the computing device calculates a usage fee for the user using the pricing formula, wherein the pricing formula calculates the usage fee based on at least one of signal transmission volume, transmission time, data access rights and communication costs.

[0015] In one embodiment of the present invention, each of the plurality of historical sensing data and the plurality of current sensing data is one or a combination of spectral data, dissolved oxygen level, pH value, water flow velocity, air temperature, water temperature, optical image, thermal image, wind speed, salinity, biological information and underwater topographic location.

[0016] In one embodiment of the present invention, each of the plurality of sensing data acquisition units is one or a combination of a multispectral sensor, a dissolved oxygen sensor, a pH sensor, a fish finder, a radar, a water flow velocity sensor, a temperature sensor, an optical sensor, a wind speed sensor, a salinity sensor, and a thermal image sensor.

[0017] In one embodiment of the present invention, the selected fishing ground information includes one or a combination of the following: fishing ground range, fishing ground coordinates, fishing ground sea temperature, fishing ground water flow velocity, fishing ground upwelling location, fishing ground geographical location, fishing ground depth, and fishing ground fish species distribution.

[0018] In one embodiment of the present invention, the coordinates of the fishing grounds are represented by one or a combination of the following coordinate systems: Cartesian Coordinate System, Polar Coordinate System, Cylindrical Coordinate System, Spherical Coordinate System, Homogeneous Coordinates System, Curvilinear Coordinates System, Generalized Coordinates System, Canonical Coordinates System, Plücker Coordinates System, Barycentric Coordinate System, Geodetic Coordinate System, and Horizontal Coordinate System.

[0019] In addition, based on at least one objective of the present invention, the present invention also provides a remote sensing method for predicting fishing grounds, comprising the following steps: performing environmental sensing through a plurality of sensing data acquisition units disposed at a plurality of locations within a selected field to generate a plurality of current sensing data at the plurality of locations; storing the plurality of current sensing data at the plurality of locations through a storage unit communicatively connected to the plurality of sensing data acquisition units; establishing a fishing ground prediction model for the selected field based on a plurality of historical sensing data stored in the storage unit through a computing device communicatively connected to the storage unit and a user's terminal device; and using the fishing ground prediction model through the computing device... The model predicts the probability model of fish distribution in a selected field within a specific time period based on multiple current sensing data from multiple locations, where the specific time period refers to the time required for the food chain evolution of the selected field; and generates and transmits information on at least one selected fishing ground within the selected field to a terminal device based on the fish distribution probability model through a computing device; wherein the learning algorithm module of the computing device determines multiple water color data from multiple locations based on multiple current sensing data and multiple historical sensing data from multiple locations, and the fishing ground prediction model predicts the probability model of fish distribution in a selected field within a specific time period based on multiple water color data from multiple locations.

[0020] In one embodiment of the present invention, the learning algorithm module has a water color data correction formula, which is used to correct a plurality of current sensing data and a plurality of historical sensing data at a plurality of locations, thereby correcting a plurality of water color data at a plurality of locations, and the fishery prediction model predicts the probability model of fish distribution in a selected field at a specific time based on the corrected plurality of water color data at a plurality of locations.

[0021] In summary, the remote sensing fishing ground prediction system and method of the present invention can provide users with information on selected fishing grounds for a specific time period in any selected field, allowing users to quickly understand the distribution probability of different fish schools near their current fishing boat location. Since it predicts the distribution probability of fish schools within a specific time period, such as providing the distribution probability of fish schools in one selected field within a week, it can ensure that the fishing ground still exists when the user arrives at the selected field within a specific time period, thus avoiding the situation where the fishing ground disappears after the user arrives at the predicted location.

Implementation Method

[0024] The examiner understands the technical features, content and advantages of the present invention and the effects it can achieve. The present invention will now be described in detail with reference to the accompanying drawings and in the form of embodiments. The drawings used are only for illustration and to assist in the description and may not be the actual proportions and precise configurations after the implementation of the present invention. Therefore, the proportions and configurations of the accompanying drawings should not be used to interpret or limit the scope of the present invention in actual implementation. This should be stated in advance.

[0025] The remote sensing fishery prediction system and method provided by the present invention can be used to sense information in any random field, and transmit the previously sensed information (i.e., historical sensing data) to the computing processing center as a computing device for processing, and establish a fishery prediction model. Through the fishery prediction model, the current sensing data (or possibly the current sensing data and historical sensing data from a certain period of time ago) can be brought into the fishery prediction model to predict the probability of fish distribution in a random field at a specific time.

[0026] Please refer to Figures 1, 2, and 3. Figure 1 is a schematic diagram of the remote sensing fishery prediction system according to an embodiment of the present invention. Figure 2 is a system block diagram of the remote sensing fishery prediction system according to an embodiment of the present invention. Figure 3 is a system block diagram of the computing device of the remote sensing fishery prediction system according to an embodiment of the present invention. As shown in Figure 1, the remote sensing fishery prediction system of the present invention includes a plurality of communication node devices 101-103, 105, a plurality of sensing data acquisition units 104, a computing processing center 106, and a terminal device 107. The communication node devices are mounted on one or a combination of artificial satellites, fishing boats, drones, vehicles, and base stations. The communication node devices include antennas (directional antennas or non-directional antennas), signal transmitters, and signal receivers. In addition, the communication node devices may selectively include a platform that can rotate in multiple directions to facilitate mounting. The present invention is not limited to the above types. In the embodiment shown in Figure 1, a plurality of communication node devices 101 are mounted on an artificial satellite, a plurality of communication node devices 102 are mounted on a drone, a plurality of communication node devices 103 are mounted on a fishing boat, and a plurality of communication node devices 105 are mounted on a base station. Referring also to Figures 2 and 3, the computing center 106 can perform large-scale data processing and includes a storage unit 1061 and a computing device 1062. A plurality of sensing data acquisition units 104 are communicatively connected to the plurality of communication node devices 101-103 and 105, and the plurality of communication node devices 101-103 and 105 are communicatively connected to the storage unit 1061. The storage unit 1061 is communicatively connected to the computing device 1062 and the terminal device 107. Furthermore, the computing device 1062 has hardware circuitry and can execute software programs to configure a data preprocessing module 10621 and a learning algorithm module 10622. The data preprocessing module 10621 and the learning algorithm module 10622 are signal-connected to each other. The data preprocessing module 10621 is used to preprocess the received data, and the learning algorithm module 10622 is used to establish a fishing ground prediction model PRED_M. Then, the fishing ground prediction model PRED_M is used to obtain the random fishing ground information in any random field D.

[0027] Further explanation: The computing center 106 may be a cloud-based big data computing center. The storage unit 1061 may be one or a combination of an electromagnetic recording device, a hard disk, and non-volatile memory (not volatile memory, such as DRAM). The computing device 1062 may be a central processing unit, memory, and a hard disk-based cloud computing device. Furthermore, the learning algorithm module 10622 can utilize backpropagation, supervised learning, semi-supervised learning, ensemble learning, active learning, reinforcement learning, generative models, discriminative models, long short-term memory, object detection, instance segmentation, and diffusion models to build the fishery prediction model PRED_M. The generative models include Gaussian Mixture Model, Maximum Likelihood Estimation, Hidden Markov Model, Naive Bayes classifier, Generative Adversarial Network, and Generative Pre-trained Transformer. The discriminative models include Logistic Regression. The learning algorithm module 10622 can perform various algorithms, including linear regression, support vector machine, decision tree, and extreme gradient boosting. However, this invention does not limit the specific algorithm used by the learning algorithm module 10622 for modeling.In addition, the method of using multiple historical sensing data to build the fishing ground prediction model PRED_M is as follows: label each set of multiple historical sensing data to correspond to the fish distribution probability model within a specific time after that time. Input the multiple sets of labeled multiple historical sensing data and the corresponding multiple fish distribution probability models into the learning algorithm module 10622. After multiple iterations of calculation, the convergence condition can be reached, and the fishing ground prediction model PRED_M can be built accordingly.

[0028] First, to predict the information of a selected fishing ground within a selected field D using a remote sensing fishing ground prediction system, it is necessary to collect historical sensing data of the locations of all multiple sensing data acquisition units 104 within the selected field D to train and establish the fishing ground prediction model PRED_M. The collection method is to transmit the historical sensing data of the locations of the multiple sensing data acquisition units 104 within the selected field D to the storage unit 1061 of the computing center 106 for storage through multiple communication node devices 101~103, 105. The historical sensing data is the past sensing data relative to the current sensing data. Next, the data preprocessing module 10621 of the computing device 1062 of the computing center 106 performs data preprocessing on the multiple historical sensing data. Then, the learning algorithm module 10622 establishes and trains the fishing ground prediction model PRED_M based on the multiple preprocessed historical sensing data.

[0029] After the fishery prediction model PRED_M is established, the fishery prediction model PRED_M can be used to predict the probability of fish distribution in the selected field D. To predict the probability of fish distribution in the selected field D, the current sensing data of the location of the multiple sensing data acquisition units 104 must be transmitted to the storage unit 1061 of the computing center 106 for storage through multiple communication node devices 101~103, 105 of the selected field D. Then, the data preprocessing module 10621 of the computing device 1062 of the computing center 106 performs data preprocessing on the multiple current sensing data. Then, the multiple preprocessed current sensing data are brought into the fishery prediction model PRED_M, and the fishery prediction model PRED_M is used to predict the probability model of fish distribution in the selected field D within a specific time period. The specific time period refers to the time required for the food chain evolution of the selected field D. Finally, based on the fish school distribution probability model, at least one random fishing ground information within the selected field D is generated, and the random fishing ground information is transmitted to the terminal device 107. Further, the fish school distribution probability model includes the probability of a fish school appearing at each position within the selected field D within a specific time period; that is, it is a probability distribution function F(x,y,z,t), where x, y, and z are position coordinates, and t is a point in time within a specific time period.

[0030] Further explanation: The sensing data acquisition unit 104 may be one or a combination of a multispectral sensor, dissolved oxygen sensor, pH sensor, fish finder, radar, water flow velocity sensor, temperature sensor, optical sensor, wind speed sensor, salinity sensor, and thermal image sensor. Historical sensing data and current sensing data may be one or a combination of spectral data, dissolved oxygen, pH, water flow velocity, air temperature, water temperature, optical image, thermal image, wind speed, salinity, biological information, and underwater topographic location. However, the types of technologies applied in this invention are not limited to those described above.

[0031] The data preprocessing module 10621 can use one or a combination of genetic algorithms, data clustering, image processing, edge detection, chi-square test, EM algorithm (Expectation-Maximization Algorithm), and data labeling for data preprocessing. In other words, data preprocessing is one or a combination of data cleaning methods, data standardization methods, and data normalization methods. Data cleaning methods address duplicate data, inconsistent data types, outliers, missing data, structurally incorrect data, and noise in the data. Data standardization methods utilize one or a combination of Z-score, mean, Max-Min algorithm, MaxAbs standardization, robust scaling, and standard deviation to process the data. Data normalization methods can include label encoding. The present invention uses one or a combination of preprocessing techniques, such as one-hot encoding and one-hot encoding, to process data. Furthermore, the present invention is not limited to the aforementioned data preprocessing methods.

[0032] Furthermore, in one embodiment, the learning algorithm module 10622 determines the water color data at different locations within the randomized field D based on a plurality of current sensing data and a plurality of historical sensing data. The learning algorithm module 10622 contains a water color data correction formula. The learning algorithm module 10622 uses this formula to correct the current and historical sensing data sensed by the plurality of sensing data acquisition units 104 within the randomized field D, thereby correcting the water color data at different locations. Then, the fishery prediction model PRED_M predicts the probability model of fish distribution in the randomized field D within a specific time period based on the corrected water color data at different locations. The presence or absence of the aforementioned water color correction formula is not a limitation of this invention. The reason for needing to correct water color data is that water color is easily affected by air medium, suspended particles, or atmospheric optics, which can influence the water color imaging obtained by the sensing device. Therefore, in some cases, it is necessary to correct the water color data.

[0033] In another embodiment, the learning algorithm module 10622 has a pricing formula. The learning algorithm module 10622 calculates the user's usage fee based on at least one of the following: signal transmission volume, transmission time, data access rights, and communication fees when the user uses the system.

[0034] Next, please refer to Figures 4A and 4B together. Figure 4A is a schematic diagram of the screen provided by the remote sensing fishery prediction system of the present invention, and Figure 4B is a schematic diagram of another screen provided by the remote sensing fishery prediction system of the present invention. The difference between Figure 4A and Figure 4B is that the user can select the predicted fishery F1 on the screen of Figure 4A, while the screen of Figure 4B displays the selected fishery information of the predicted fishery F1. In the embodiments of Figures 4A and 4B, the user wants to use the telemetry fishing ground prediction system of the present invention to predict the probability of fish distribution in the vicinity of the user's selected current location (e.g., the current location of the user's fishing boat, or the location where the fishing boat will arrive soon). The telemetry fishing ground prediction system uses multiple communication node devices 101-103, 105 to transmit multiple current sensing data from multiple sensing data acquisition units 104 in the vicinity of the user's selected current location to the computing processing center 106. Then, the computing device 1062 of the computing processing center 106 uses the fishing ground prediction model PRED_M to predict the probability of fish distribution in the vicinity of the user's current location (or the current location). (Based on the sensor data and historical sensor data over a certain period of time), a probability model of fish distribution in the vicinity of the user's current location is predicted within a specific time period. Then, based on the probability model of fish distribution, random fishing ground information is generated and sent to the user's terminal device 107. At this time, the user can see all the random fishing ground information and predicted fishing grounds FL1~FL4 near the user's current location BL through the terminal device 107. The random fishing ground information will be presented on the user's terminal device 107 in one or a combination of the following: fishing ground range, fishing ground coordinates, fishing ground sea temperature, fishing ground water flow velocity, fishing ground upwelling location, fishing ground geographical location, fishing ground depth, and fishing ground fish species distribution.

[0035] Further explanation: The user's terminal device 107 may be one or a combination of a mobile phone, tablet computer, laptop computer, personal computer, and quantum computer, but this invention is not limited thereto. The representation of the fishing ground coordinates for the selected fishing ground information may be one or a combination of the Cartesian Coordinate System, Polar Coordinate System, Cylindrical Coordinate System, Spherical Coordinate System, Homogeneous Coordinates System, Curvilinear Coordinates System, Generalized Coordinates System, Canonical Coordinates System, Plücker Coordinates System, Barycentric Coordinate System, Geodetic Coordinate System, and Horizontal Coordinate System, but this invention is not limited thereto.

[0036] Next, please refer to Figure 5, which is a flowchart of the remote sensing fishery prediction method according to an embodiment of the present invention. Each step of the remote sensing fishery prediction method of the present invention is executed by the corresponding element of the aforementioned remote sensing fishery prediction system. To achieve the goal of predicting the probability of fish distribution in a selected field by acquiring multiple current sensing data of the positions of multiple sensing data acquisition units within the selected field, a fishery prediction model must first be established. Therefore, in step S501, the remote sensing fishery prediction system obtains multiple historical sensing data of different positions within the selected field from the storage unit through the computing device, and establishes a fishery prediction model based on the multiple historical sensing data. After the fishery prediction model is established, in step S502, the remote sensing fishery prediction system acquires multiple current sensing data through the sensing data acquisition units of multiple different positions within the selected field. In step S503, the telemetry fishery prediction system transmits multiple current sensing data points to the storage unit of the computing center via multiple different communication node devices for storage. In step S504, the computing device uses the fishery prediction model and the multiple current sensing data points at multiple locations stored in the storage unit to predict the probability model of fish distribution in a selected field within a specific time period, where the specific time period refers to the time required for the food chain evolution of the selected field.

[0037] It is worth noting that a fish farm typically forms when a group of algae or plankton containing chlorophyll first exists in a certain area. After several days or a week, the algae or plankton grow and attract fish to feed, thus forming a fish farm in that area. Therefore, a specific time is related to the time required for the food chain evolution of the selected area. The algae or plankton containing chlorophyll can obtain chlorophyll-related sensing data through the sensing data acquisition unit. Then, the learning algorithm module determines the water color data of this area based on the sensing data. The water color data represents the aggregation status of the algae or plankton, and the aggregation status of the algae or plankton affects the probability of fish farm formation. Therefore, the fish farm prediction model can predict the fish distribution probability model based on the multiple water color data corresponding to the multiple current sensing data. Next, referring to step S505, the computing device will generate at least one selected fish farm information based on the fish distribution probability model and transmit the selected fish farm information to the user's terminal device for the user to view.

[0038] As can be seen from the above description of the implementation, compared with the prior art and products, the present invention has the following advantages:

[0039] 1. The remote sensing fishing ground prediction system and method of the present invention can provide users with information on fishing grounds within a specific time period for any selected field, such as providing the probability of fish distribution within a week within a 20-nautical-mile range of the user's current location, so that users can quickly understand the probability of different fish distribution near their current location.

[0040] 2. The remote sensing fishery prediction system and method of the present invention predicts the probability of fish distribution based on pre-sensing data. Therefore, the predicted probability range of fish distribution is dynamic information, allowing users to obtain the most accurate prediction information.

[0041] 3. The remote sensing fishing ground prediction system and method of the present invention has the function of predicting the probability of future fish distribution in a selected field. Therefore, it can predict the fish distribution after the user's boat arrives at the predicted location, thus avoiding the situation where the fishing ground disappears after the user arrives at the predicted location.

[0042] The above-described embodiments are only for illustrating the technical ideas and features of the present invention. Their purpose is to enable those skilled in the art to understand the content of the present invention and implement it accordingly. They should not be used to limit the patent scope of the present invention. That is, all equivalent changes or modifications made in accordance with the spirit disclosed in the present invention should still be covered within the patent scope of the present invention. [Simplified Explanation of the Diagram]

[0022] The accompanying drawings are provided to enable those skilled in the art to further understand the invention, and are incorporated in and constitute a part of the specification of the invention. The drawings illustrate exemplary embodiments of the invention and are used together with the specification of the invention to explain the principles of the invention.

[0023] Figure 1 is a schematic diagram of the remote sensing fishery prediction system according to an embodiment of the present invention. Figure 2 is a system block diagram of the remote sensing fishery prediction system according to an embodiment of the present invention. Figure 3 is a system block diagram of the computing device of the remote sensing fishery prediction system according to an embodiment of the present invention. Figure 4A is a schematic diagram of a screen provided by the remote sensing fishery prediction system according to an embodiment of the present invention. Figure 4B is a schematic diagram of another screen provided by the remote sensing fishery prediction system according to an embodiment of the present invention. Figure 5 is a flowchart of the remote sensing fishery prediction method according to an embodiment of the present invention.

Claims

1. A remote sensing fishing ground prediction system for transmitting information about at least one selected fishing ground in a selected field (D) to a user via a terminal device (107), comprising: a plurality of sensing data acquisition units (104) disposed at a plurality of locations in the selected field (D) for performing environmental sensing to generate a plurality of current sensing data at the plurality of locations; A storage unit (1061), communicatively connected to the plurality of sensing data acquisition units (104), is used to store the plurality of current sensing data and the plurality of historical sensing data for the plurality of locations; and a computing device (1062), communicatively connected to the storage unit (1061) and the terminal device (107), wherein the computing device (1062) runs a software program to configure a data preprocessing module (10621) and a learning algorithm module (10622), wherein the learning algorithm module (10622) uses reinforcement learning. The system employs a learning technique and establishes a fishery prediction model (PRED_M) for the selected field (D) based on a plurality of historical sensing data. It also determines a plurality of water color data for a plurality of locations based on a plurality of current sensing data and a plurality of historical sensing data. The fishery prediction model (PRED_M) is then used to predict a probability model of fish population distribution in the selected field (D) within a specific time period based on the plurality of current sensing data and the plurality of water color data at the plurality of locations. The system then generates the selected fishery information for the terminal device (107) based on the fish population distribution probability model. The specific time period is related to the time required for the evolution of a food chain in the selected field (D). The fish population distribution probability model is a probability distribution function F(x,y,z,t), where x, y, and z are location coordinates, and t is a point in time within a specific time period.

2. The telemetry fishery prediction system as claimed in claim 1, wherein in a prediction mode, the data preprocessing module (10621) is used to preprocess the plurality of current sensing data, and the learning algorithm module (10622) uses the fishery prediction model (PRED_M) to predict the probability model of the fish distribution in the random field (D) at the specific time based on the preprocessed plurality of current sensing data.

3. The telemetry fishing ground prediction system as claimed in claim 1, wherein in a training mode, the data preprocessing module (10621) is used to preprocess the plurality of historical sensing data, and the learning algorithm module (10622) establishes the fishing ground prediction model (PRED_M) based on the plurality of historical sensing data.

4. The telemetry fishing ground prediction system as described in claim 2, wherein the data preprocessing module (10621) uses one or a combination of genetic algorithms, data clustering, image processing, edge detection, chi-square test, Expectation-Maximization Algorithm, and data labeling for data preprocessing.

5. The telemetry fishing ground prediction system as claimed in claim 1 further comprises: a plurality of communication node devices (101, 102, 103, 105) communicatively connected to the plurality of sensing data acquisition units (104) and the storage unit (1061), wherein each of the communication node devices (101, 102, 103, 105) is mounted on one of an artificial satellite, a fishing boat, a drone and a vehicle.

6. The remote sensing fishery prediction system as claimed in claim 1, wherein each of the plurality of historical sensing data and the plurality of current sensing data is one or a combination of spectral data, dissolved oxygen level, pH value, water flow velocity, air temperature, water temperature, optical image, thermal image, wind speed, salinity, biological information and underwater topographic location.

7. The telemetry fishing ground prediction system as claimed in claim 1, wherein each of the plurality of sensing data acquisition units (104) is one or a combination of a multispectral sensor, a dissolved oxygen sensor, a pH sensor, a fish finder, a radar, a water flow velocity sensor, a temperature sensor, an optical sensor, a wind speed sensor, a salinity sensor and a thermal image sensor.

8. The telemetry fishing ground prediction system as described in claim 1, wherein the selected fishing ground information includes one or a combination of a fishing ground range, a fishing ground coordinate, a fishing ground sea temperature, a fishing ground current velocity, a fishing ground upwelling location, a fishing ground geographical location, a fishing ground depth, and a fishing ground fish species distribution.

9. The remote sensing fishing ground prediction system as described in claim 8, wherein the coordinates of the fishing ground are represented by one or a combination of the following coordinate systems: Cartesian Coordinate System, Polar Coordinate System, Cylindrical Coordinate System, Spherical Coordinate System, Homogeneous Coordinates System, Curvilinear Coordinates System, Generalized Coordinates System, Canonical Coordinates System, Plücker Coordinates System, Barycentric Coordinate System, Geodetic Coordinate System, and Horizontal Coordinate System.

10. A method for predicting fishing grounds using telemetry, employing the telemetry fishing ground prediction system as described in any one of claims 1 to 9, comprising: performing environmental sensing through a plurality of sensing data acquisition units (104) located at a plurality of locations within a randomized field (D) to generate a plurality of current sensing data for the plurality of locations; storing the plurality of current sensing data for the plurality of locations through a storage unit (1061) communicatively connected to the plurality of sensing data acquisition units (104); and establishing a fishing ground prediction model (PRED_M) for the randomized field (D) based on the plurality of historical sensing data stored in the storage unit (1061) through a computing device (1062) communicatively connected to the storage unit (1061) and a terminal device (107) of a user. The computing device (1062) uses the fishery prediction model (PRED_M) to predict the probability model of fish distribution in the random field (D) within a specific time period based on the multiple current sensing data and the multiple water color data at the multiple locations. The probability model of fish distribution is a probability distribution function F(x,y,z,t), where x, y, and z are location coordinates, t is a point in time within the specific time period, and the specific time is related to the time required for the evolution of a food chain in the random field (D). The computing device (1062) generates and transmits information on at least one random fishery in the random field (D) to the terminal device (107) based on the probability model of fish distribution. The computing device (1062) runs a software program to configure a data preprocessing module (10621) and a learning algorithm module (10622). The learning algorithm module (10622) uses reinforcement learning technology to build the fishery prediction model (PRED_M) and determines the multiple water color data at the multiple locations based on the multiple current sensing data and the multiple historical sensing data at the multiple locations.

Citation Information

Patent Citations

  • Fishery resource prediction method and device, storage medium and electronic equipment

    CN114444819A

  • Fish condition visualization method and related equipment

    CN115294463A

  • Method for position, auction and distribution of fishery harvest

    TW202205188A