Fishing state determination method and system

Abstract

To provide a fishing state determination method and system.SOLUTION: A fishing state determination method includes: performing model parameter training on a classification model using historical motion state data set of a fishing vessel under different casting postures; acquiring current motion state data of the fishing vessel during casting; inputting the current motion state data into the trained classification model to carry out casting posture prediction, and outputting a probability corresponding to each casting posture; and determining the casting posture corresponding to the maximum probability as the current casting posture. By intelligently calculating the motion state data of the fishing vessel during casting and accurately outputting the casting posture, the system enables a user to sense whether or not the casting posture is correct and quickly grasp various casting postures in an early stage of learning, and can reduce problems such as casting errors, unnecessary exertion during casting, and a backlash caused by the casting errors.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to the field of fishing data processing, and more particularly to a fishing state determination method and system. [Background technology]

[0002] With the boom of outdoor fishing, more and more people are participating in fishing. During the fishing process, users face many fishing operation problems, such as not being able to cast accurately, frequent tangled fishing lines, excessive physical exhaustion, and difficulty in learning certain fishing movements.

[0003] To improve a user's fishing operation ability, especially the ability to operate in various casting positions, an accurate casting position is the basis for effective casting. However, currently, learning various casting positions is often done through short video lessons or on-site lessons by real people, which requires high trial and error costs and slow progress. Summary of the Invention [Problem to be solved by the invention]

[0004] The objective of the present invention is to provide a fishing state determination method and system to solve the problem of how to help users improve their operation for various casting positions. [Means for solving the problem]

[0005] According to a first aspect, an embodiment of the present invention provides a fishing state determination method, comprising: training model parameters for the classification model using a historical motion state dataset of the fishing vessel at different casting attitudes; A step of acquiring current motion state data of the fishing boat during casting; inputting the current motion state data into the trained classification model to perform casting posture prediction, and outputting a probability corresponding to each casting posture; determining the casting position corresponding to the highest probability as the current casting position.

[0006] According to a second aspect, an embodiment of the present invention provides a fishing state determination system, comprising: a model training unit for performing model parameter training on the classification model using a historical motion state dataset of the fishing vessel at different casting attitudes; a data acquisition unit for acquiring current motion state data of the fishing boat during casting; a posture prediction unit for inputting the current motion state data into the trained classification model to perform casting posture prediction and outputting a probability corresponding to each casting posture; a posture determining unit for determining the casting posture corresponding to the maximum probability as the current casting posture. [Effects of the Invention]

[0007] The beneficial effects of the present invention are as follows: By intelligently calculating the motion state data of the fishing boat during casting and accurately outputting the current casting position, it is possible to help the user determine whether the casting position is correct, and it is possible to quickly grasp various casting positions at the beginning of learning, thereby reducing problems such as casting errors, unnecessary physical exhaustion, and line tangles due to casting errors. [Brief explanation of the drawings]

[0008] In order to more clearly describe the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings that need to be used in the description of the embodiments. Obviously, the drawings in the following description are some embodiments of the present invention, and those skilled in the art can derive other drawings based on these drawings without any creative efforts.

[0009] [Figure 1] 1 is a flowchart of a fishing state determination method provided in an embodiment of the present invention. [Figure 2] 1 is a sub-flowchart of a fishing state determination method provided in an embodiment of the present invention. [Figure 3] FIG. 10 is yet another sub-schematic diagram of the fishing state determination method provided in an embodiment of the present invention. [Figure 4] FIG. 10 is yet another sub-schematic diagram of the fishing state determination method provided in an embodiment of the present invention. [Figure 5] FIG. 10 is yet another sub-schematic diagram of the fishing state determination method provided in an embodiment of the present invention. [Figure 6] FIG. 10 is yet another sub-schematic diagram of the fishing state determination method provided in an embodiment of the present invention. [Figure 7] FIG. 10 is yet another sub-schematic diagram of the fishing state determination method provided in an embodiment of the present invention. [Figure 8] FIG. 10 is yet another sub-schematic diagram of the fishing state determination method provided in an embodiment of the present invention. [Figure 9] FIG. 10 is yet another sub-schematic diagram of the fishing state determination method provided in an embodiment of the present invention. [Figure 10] 1 is a schematic block diagram of a fishing state determination system provided in an embodiment of the present invention. FIG. [Figure 11] 1 is a schematic diagram of an excerpt from three-axis angular acceleration data provided in an embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0010] The technical solutions in the embodiments of the present invention will be described below clearly and completely with reference to the drawings in the embodiments of the present invention, but it is clear that the described embodiments are only some of the embodiments of the present invention and not all of the embodiments. All other embodiments made by those skilled in the art based on the embodiments of the present invention without creative efforts fall within the protection scope of the present invention.

[0011] It should be understood that when used in this specification and the appended claims, the terms "comprises" and "including" indicate the presence of stated features, wholes, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or groups thereof.

[0012] It should also be understood that the terminology used in the specification of the present invention is for the purpose of describing particular embodiments only, and is not intended to be limiting of the present invention. As used in the specification of the present invention and the appended claims, the singular forms "a," "one," and "the" are intended to include the plural forms unless the context clearly dictates otherwise.

[0013] It is to be further understood that the term "and / or" as used in the present specification and the appended claims refers to and includes any and all possible combinations of one or more of the associated listed items.

[0014] Please refer to FIG. 1, which is a flowchart of a fishing state determination method provided in an embodiment of the present invention.

[0015] As shown in FIG. 1, the method includes steps S101 to S104.

[0016] In S101, model parameter training is performed on the classification model using historical motion state data sets of the fishing vessel at different casting attitudes.

[0017] In this step, historical motion state data of the fishing vessel at each casting position is collected and constructed as a historical motion state dataset. The historical motion state dataset includes previously collected motion data generated when the fishing vessel casts at different casting positions. The different casting positions can be classified into multiple types, including over-the-shoulder throwing, ultra-long distance throwing, side throwing, swing throwing, water-skimming, and ejection throwing. The historical motion state dataset is divided into a training set, a test set, and a validation set, and model parameter training is performed for the classification model.

[0018] In S102, current motion state data of the fishing boat at the time of casting is acquired.

[0019] In this step, the motion sensor installed on the fishing vessel can collect motion data generated throughout the entire current casting process of the fishing vessel. The entire casting process is divided into a casting preparation stage, an actual casting stage, a casting completion stage, and a reeling stage. Here, the fishing vessel is relatively stationary during the casting preparation stage, the casting completion stage, and the reeling stage, but the fishing vessel moves rapidly during the actual casting stage. Therefore, after collecting motion data generated throughout the entire casting process, the motion data of the fishing vessel during the actual casting stage is selected as current motion state data during the current casting.

[0020] In S103, the current motion state data is input into the trained classification model to perform casting posture prediction, and the probability corresponding to each casting posture is output.

[0021] In S104, the casting position corresponding to the maximum probability is determined as the current casting position.

[0022] In steps S103-S104, since the operations such as casting angle and force are different in different casting postures, the motion state data of different casting postures have different regularities. Therefore, by inputting the current motion state data collected in real time during casting into the trained classification model to perform casting posture prediction, the probability of each casting posture corresponding to the current motion state data can be accurately output, and the casting posture corresponding to the maximum probability is determined as the current casting posture, thereby realizing accurate casting posture recognition.

[0023] In this embodiment, the current motion state data of the fishing boat during casting is intelligently calculated, and then combined with the theoretical analysis of the casting motion, the current casting posture is output, thereby helping the user to sense whether the casting posture is correct or not, allowing the user to quickly grasp various casting postures at the early stage of learning, and reducing problems such as casting errors, unnecessary physical exhaustion, and line tangles caused by casting errors.

[0024] In some specific embodiments of the classification model, an XGBoost model can be selected and model parameter training can be performed on the XGBoost model using a historical motion data set. Specifically, the model parameters of the XGBoost model include the learning target, the number of classification targets, the minimum loss function, the maximum single-tree depth, the L2 weight regularization term, the sampling rate per tree, and the occupancy ratio of the number of columns controlling sampling. Here, the number of classification targets refers to the types of casting postures in this application, i.e., the currently set over-the-shoulder throw, ultra-long throw, side throw, swing throw, water skip, and injection throw. A non-casting posture can also be added, i.e., the final number of classification targets is the currently set number of casting postures plus 1. Based on the trained XGBoost model, the current motion state data of the fishing boat during casting can be input into the XGBoost model, and the probabilities of each currently set casting posture and non-casting posture can be output. The classification model in this embodiment can also be a model such as a LightGBM model or a CatBoost model, which can be iteratively trained using a classifier.

[0025] In one embodiment, as shown in FIG. 2, step S102 includes: S201 collects triaxial angular acceleration data of the fishing boat during the casting process; S202 extracts motion state data from the three-axis angular acceleration data, the motion state data including a positive angular change, a negative angular change, a maximum angular acceleration, and a minimum angular acceleration of the y-axis, and a positive angular change, a negative angular change, a maximum angular acceleration, and a minimum angular acceleration of the z-axis.

[0026] In this embodiment, during the casting process, the motion sensor collects three-axis angular acceleration data of the fishing boat during the casting process, i.e., collects data of gyro_x, gyro_y and gyro_z during the casting process. Gyro_x is the orientation of the fishing rod itself and does not change significantly during the casting process, but gyro_y and gyro_z change significantly during the casting process. The faster the casting motion, the larger the instantaneous values ​​of gyro_y and gyro_z. Therefore, the "moving state" and the "non-moving state" are distinguished based on the magnitude of the instantaneous values ​​of gyro_y and gyro_z.

[0027] To facilitate understanding, a schematic diagram of an excerpt from three-axis angular acceleration data is shown in Figure 11, where the horizontal axis represents time in units of 1 / 25 seconds, and the vertical axis represents the original value after gyro sensor conversion. The dotted line gyro_x shows large fluctuations, while the dotted lines gyro_y and gyro_z show more regular changes. Therefore, the magnitude of the instantaneous values ​​of gyro_y and gyro_z is used to distinguish between motion state data and non-motion state data in the three-axis angular acceleration data. The motion dotted line represents the criterion for whether or not it is motion state data, and a preset motion state value (e.g., 1000) is used to represent the boundary between motion state data and non-motion state data. In other words, motion state data begins at the position where the absolute value of the instantaneous values ​​of gyro_y and gyro_z is greater than the initially preset motion state value, and continues until the position where the instantaneous values ​​of gyro_y and gyro_z are smaller than the preset motion state value. Data in this positional interval is motion state data, and conversely, data outside this positional interval is non-motion state data.

[0028] In the motion state data of the three-axis angular acceleration data, a positive angular change on the y-axis refers to the addition of consecutive gyro_y values ​​> 0, a negative angular change on the y-axis refers to the addition of consecutive gyro_y values ​​< 0, the maximum angular acceleration on the y-axis refers to the maximum value of the gyro_y data segment, the minimum angular acceleration on the y-axis refers to the minimum value of the gyro_y data segment, a positive angular change on the z-axis refers to the addition of consecutive gyro_z values ​​> 0, a negative angular change on the z-axis refers to the addition of consecutive gyro_z values ​​< 0, the maximum angular acceleration on the z-axis refers to the maximum value of the gyro_z data segment, and the minimum angular acceleration on the z-axis refers to the minimum value of the gyro_z data segment.

[0029] In some embodiments, during actual casting, errors will occur in the three-axis angular acceleration data collected by the motion sensor, so the collected data can be filtered after collection. The filtering method can be first-order smoothing filtering, with filtering parameters Q and W, the sum of Q and W is 1, and the filtering formula is current filtered value = Q * previous original value + W * current original value, which can improve the accuracy of the collected data.

[0030] In one embodiment, as shown in FIG. 3, step S103 includes: S301 inputs motion state data of the three-axis angular acceleration data into a trained classification model to predict casting postures and output scores corresponding to each casting posture; and S302, which performs probability conversion based on the score of each casting position and the total score of all the casting positions, and outputs a probability corresponding to each casting position.

[0031] In this embodiment, the motion state data of the three-axis angular acceleration data is input into the classification model after training, and combined with the model parameters of the classification model to calculate the score of each casting posture corresponding to the current motion state data.The score of each casting posture is then substituted into an activation function (exponential function) for conversion, and the probability of each casting posture is then calculated using a probability calculation formula.

[0032] Specifically, the probability calculation formula is p(n)=s(n) / sum(s(0)+...+s(N)), where n represents the nth casting position, N represents the number of all casting positions, p(n) represents the probability of the nth casting position, and s(n) represents the score of the nth casting position.

[0033] Based on this, the probability of each casting position is calculated, and then the casting position with the highest probability is selected as the current casting position.

[0034] In some other embodiments, in order to improve the accuracy of recognizing the current casting posture, a preset threshold can be added, and the casting posture with the highest probability that reaches the preset threshold is selected as the current casting posture; that is, if there is a casting posture with the highest probability but that does not reach the preset threshold, it is possible that the casting posture with the highest probability is not the correct current casting posture; in other words, it is determined that the current casting posture is not standard, and the current casting posture cannot be accurately recognized, thereby making it easy for the user to make changes to improve subsequent casting accuracy.

[0035] In one embodiment, as shown in FIG. 4, the fishing state determination method of the present invention further comprises: S401, which presets the brake mode corresponding to each casting posture; After determining that the current casting action has ended, S402 automatically executes a brake mode corresponding to the current casting posture, and the method for determining that the current casting action has ended includes S402, which is a method for determining that the current casting action has ended when the collected three-axis angular gravity acceleration data switches from a moving state to a non-moving state for the first time.

[0036] In this embodiment, after the casting motion is completed, the fishing boat is usually braked to avoid problems such as line tangles. In order to accommodate stable reeling in different casting positions, different brake modes pre-set for different casting positions are used. Specifically, when the data in the three-axis angular gravity acceleration data switches from a moving state to a non-moving state for the first time, the current casting motion is completed, and the current casting position is also output according to the classification model. The fishing boat is also basically in a stationary environment, but the casting motion is completed but the fishing boat is still reeling out. At this time, the fishing boat can automatically execute the brake mode corresponding to the current casting position to prevent line tangles.

[0037] In one embodiment, as shown in FIG. 5, the fishing state determination method of the present invention further comprises: S501 collecting current acceleration data in a current casting posture, the current acceleration data including current three-axis gravitational acceleration data and current two-axis angular acceleration data, the current two-axis angular acceleration data including current y-axis angular acceleration data and current z-axis angular acceleration data; S502: performing a feature difference comparison between the current acceleration data and preset standard acceleration data to obtain a feature score of the current acceleration data; and S503 presenting a lesson based on the feature score of the current acceleration data.

[0038] In this embodiment, a motion sensor is used to collect current acceleration data of the fishing boat during the casting process, i.e., current triaxial gravitational acceleration data (Acc_x, Acc_y, and Acc_z) and current biaxial angular acceleration data (gyro_y and gyro_z). The collected current acceleration data is compared with pre-constructed standard acceleration data generated by casting in the same position, and a feature score for the current acceleration data is obtained. The level of the feature score can reflect problems in the current casting process, allowing for targeted training suggestions to be made for the casting process. The collected current acceleration data can be filtered using first-order smoothing filtering to improve the accuracy of the score.

[0039] In one embodiment, as shown in FIG. 6, step S502 includes: S601: extracting current motion state data from current acceleration data, non-motion state data within a set time length before and after the current motion state, and data at the current line winding stage, to obtain a plurality of current time series data and construct them as a current comparison time series data set; S602: performing a feature difference comparison between each current time series data in the current comparison target time series data set and corresponding standard time series data in the standard acceleration data, and outputting a feature score of the corresponding item of each current time series data; and S603, performing a weighted addition calculation on the feature scores of the same item of all the current time-series data in the currently compared dataset, and outputting a final score for each feature of the currently compared dataset.

[0040] In this embodiment, the current acceleration data may be data collected throughout the entire casting process, and current motion state data, non-motion state data within a set time length before and after the current motion state, and current line winding stage data are extracted from the current three-axis gravitational acceleration data, and current motion state data, non-motion state data within a set time length before and after the current motion state, and current line winding stage data are extracted from the current two-axis angular acceleration data, thereby obtaining multiple current time series data and constructing a current comparison time series data set.

[0041] In this embodiment, in order to ensure data consistency during feature difference comparison, an interpolation method is used to unify the data lengths of the currently compared time-series data set and the standard acceleration data. For example, of the current motion state data in the current acceleration data and the corresponding standard motion state data in the standard acceleration data, the longer data of the two is used as the standard, and the shorter data of the two is interpolated to ensure consistency with the longer data.

[0042] In this embodiment, the feature difference comparison process involves comparing the collected current time series data corresponding to Acc_x, Acc_y, Acc_z, gyro_y, and gyro_z with preset standard time series data in a one-to-one correspondence, where the features include at least one of variance, standard deviation, growth rate, change rate, slope, amplitude, correlation coefficient, and covariance. A standard score for each feature is set based on the number of feature items, i.e., the preset total score divided by the number of feature items equals the standard score for each feature.

[0043] Specifically, the current motion state data, the non-motion state data within a set time length before and after the current motion state, and the data of the current yarn winding stage are divided into three data categories. Time series data are collected for each data category in chronological order, and five current time series data, i.e., Acc_x, Acc_y, Acc_z, gyro_y, and gyro_z, are collected for each data category. That is, the total number of current time series data in the currently compared time series dataset is 3×5=15. Correspondingly, these 15 current time series data correspond to the 15 preset standard time series data. That is, after feature comparison, a total of 15×I=15I (I is the number of feature items) feature scores are obtained. The feature scores of the same item in these 15 current time series data are weighted and added to obtain I final scores for I features of these 15 current time series data, that is, I final scores for I features of the currently compared dataset. Finally, the sum of the I final scores is calculated as the user's current overall score.

[0044] More specifically, the score weights of the 15 current time series data are set as shown in Table 1 below.

[0045] [Table 1]

[0046] Here, a+b+c=1, a1+a2+a3+a4+a5=a, b1+b2+b3+b4+b5=b, and c1+c2+c3+c4+c5=c. Based on the weighting settings in Table 1, for example, after comparing the amplitude features of the current Acc_x and the corresponding standard Acc_x, the amplitude difference between them is obtained. The difference between the amplitude feature standard score and the amplitude difference is then calculated and weighted according to the weighting a1 set in Table 1 to obtain the amplitude score of the current Acc_x. Based on this, amplitude scores for all current time series data are calculated in the same manner, and then the sum of all amplitude scores is calculated to obtain the final score for the amplitude features of the currently compared data set. Finally, based on the calculation method for the final scores of the amplitude features, final scores for the variance feature, standard deviation feature, growth rate feature, rate of change feature, slope feature, correlation coefficient feature, and covariance feature can be calculated.

[0047] Furthermore, based on the final scores of all features in the currently compared dataset, a training method can be provided for features with low scores. For example, if the final score of the amplitude feature is low, a training method can be provided to inform the user that the casting intensity is too strong, which is likely to drain stamina, and to weaken the casting intensity.

[0048] In one embodiment, as shown in FIG. 7, the fishing state determination method of the present invention further comprises: S701 collects the number of casts for each casting position; When the number of times of casting in the current casting posture reaches a predetermined number, an average value of each characteristic score in the current casting posture for the predetermined number of times is calculated (S702); and S703, determining the difference between the average value of each characteristic score and the corresponding standard score as the characteristic score compensation for the current casting posture in the subsequent casting.

[0049] In this embodiment, 15 feature scores are recorded for each casting pose. After the current casting pose has been cast a preset number of times, the average of the 15 feature scores for the preset number of times is calculated. The difference between the average and the standard score represents the difference between the actual user's casting habits and those of a professional user. Therefore, if the final score for a feature is still low after the current casting pose has been cast a preset number of times, it is considered that the actual user has personal habits. Therefore, the difference between the average of each feature score and the corresponding standard score is used as a correction score. When the user performs subsequent casting scores, the final score is corrected with the post-correction score to allow self-learning to adapt to different user habits. The correction method is: final score = original feature score + correction score.

[0050] In one embodiment, as shown in FIG. 8, the fishing state determination method of the present invention further comprises: S801, which collects the current reeling length, the current reeling speed, and the current reeling acceleration of the fishing boat in the current casting position; S802: Obtaining the maximum remaining unwinding length by fitting based on the speed change rule of the current unwinding speed and the current unwinding acceleration; S803: if the sum of the current unwinding length and the maximum remaining unwinding length is greater than the currently required set unwinding length, execute a brake operation corresponding to the magnitude of the brake force based on the set speed range in which the current unwinding speed is located; The method includes S804, which recalculates whether the sum of the current unwinding length and the maximum remaining unwinding length exceeds the currently required set unwinding length each time a brake operation is performed, until the sum of the current unwinding length and the maximum remaining unwinding length approaches the currently required set unwinding length.

[0051] In this embodiment, generally, the reeling speed due to an external force of 0 is slowly slowed down by air resistance before falling into the water, so based on the change in the current reeling speed, in combination with the type of casting posture, the maximum remaining reeling length from the current point in time until falling into the water can be predicted.

[0052] Specifically, the maximum remaining reeling length can be obtained by fitting a multivariate linear regression model that combines two parameters, the current reeling speed and the current reeling acceleration, using different model parameters for different casting postures. The specific calculation method is: Maximum remaining reeling length = a * current reeling speed + b * current reeling acceleration + c, where a, b, and c are the training parameters of the multivariate linear regression model.

[0053]

[0004] Furthermore, when the sum of the maximum remaining reeling length and the current reeling length, combined with the current reeling length, is greater than the user-required reeling length, the fishing boat's electromagnetic brake system is triggered to perform a point brake operation, with the magnitude of the braking force related to the current reeling speed; i.e., the higher the current reeling speed, the greater the braking force. In one specific embodiment, multiple set speed ranges can be set, each set speed range corresponding to a brake position, with the higher the brake position, the greater the corresponding braking force. That is, once the set speed range in which the current reeling speed is located is determined, the brake position to be performed is identified, and a braking operation corresponding to the magnitude of the braking force is performed. Specifically, the fishing boat brake principle is as follows: when casting, electromagnetic induction occurs between a magnet installed in the wire cup of the fishing boat and a coil installed in the side cover of the fishing boat. When the coil module installed in the side cover is closed, current flows in the coil module, generating a magnetic field opposite to the magnet, which prevents the wire cup from rotating. When the coil module is open, no current flows, no magnetic field is generated, and the wire cup does not rotate. Therefore, by accessing the MOS transistor connected to the coil module and inputting a constant brake frequency to the MOS transistor, the MOS transistor can be controlled to turn on and off, thereby opening and closing the coil module. In other words, different brake frequency duty ratios can be set for different brake positions; the higher the brake position, the greater the duty ratio of the corresponding brake frequency, and the greater the resistance to unwinding the wire cup.

[0054] Furthermore, each time the brake is operated, the maximum remaining unwinding length after braking is recalculated based on the above-mentioned method for calculating the maximum remaining unwinding length, and whether the sum of the current unwinding length and the maximum remaining unwinding length exceeds the currently required set unwinding length is recalculated until the sum of the current unwinding length and the maximum remaining unwinding length approaches the currently required set unwinding length (i.e., the difference is less than the set length).

[0055] In one embodiment, as shown in FIG. 9, the fishing state determination method of the present invention further comprises: The S901 calculates the resting calories burned by the user throughout the fishing process, Obtain the number of castings in the entire fishing process, the casting posture of each casting, and the gravity acceleration amplitude of each casting, and calculate the casting calories generated in the entire fishing process according to a preset casting calorie consumption table (S902), where the casting calorie consumption table includes calories consumed by different casting postures under different gravity acceleration amplitudes; S903: obtain the total reeling length of the fishing boat during the entire fishing process, and input the total reeling length into a preset first regression model to output the reeling calorie required to reel in the total reeling length; S904: obtain the user's walking steps and walking frequency during the entire fishing process by the fishing boat, and input the walking steps and walking frequency into a preset second regression model to output the corresponding walking calories; S905 for calculating the sum of the resting calories, the casting calories, the reeling calories and the walking calories to obtain the fishing calories consumed by the user during the entire fishing process.

[0056] In this embodiment, the resting calories generated by the user throughout the fishing process can be calculated using existing standard resting calorie calculation methods based on the received user data.

[0057] In this embodiment, the casting calories burned during a user's fishing process are mainly due to the work of casting with the arms. Therefore, the casting calories can be classified according to the casting posture and the magnitude of the gravitational acceleration amplitude during casting. The acc width can be divided into multiple ranges in ascending order and combined with all casting postures to classify into multiple scenes. Based on this, data on multiple scenes of an expert user is pre-constructed, and the average casting calories burned during casting for each scene is calculated, thereby obtaining reference values ​​for the multiple scenes. In this way, the casting calories burned during the user's fishing process can be calculated based on the corresponding reference values, thereby obtaining the casting calories generated during the entire fishing process.

[0058] In this embodiment, a univariate linear regression model can be designed for the reeling calories based on the total reeling length of the fishing boat. The longer the total reeling length, the greater the corresponding reeling calories. The specific calculation method is reeling calories = a * total reeling length + b, where a and b are univariate linear regression model parameters. The model parameters are obtained by training based on the calorie data of professional equipment collected by professional users during the casting and reeling process.

[0059] In this embodiment, the walking calories are calculated by detecting whether the user is walking based on a motion sensor, extracting the number of steps from the motion sensor, and calculating the walking calories based on the number of steps and walking frequency. The specific calculation method is: walking calories = a * walking steps + b * walking frequency + c, where a, b, and c are multivariate linear regression model parameters, and the model parameters are obtained by training based on calorie data collected by professional equipment from professional users' walking processes.

[0060] The embodiment of the present invention further provides a fishing state determination system for implementing any of the embodiments of the fishing state determination method described above. Specifically, referring to Fig. 10, Fig. 10 is a schematic block diagram of the fishing state determination system provided in the embodiment of the present invention.

[0061] As shown in FIG. 10, the fishing state determination system 1000 includes: a model training unit 1001 for performing model parameter training for the classification model using historical motion state datasets of the fishing vessel at different casting attitudes; a data acquisition unit 1002 for acquiring current motion state data of the fishing boat during casting; a pose prediction unit 1003 for inputting the current motion state data into the trained classification model to perform casting pose prediction and outputting a probability corresponding to each casting pose; a posture determining unit 1004 for determining the casting posture corresponding to the maximum probability as the current casting posture.

[0062] The system intelligently calculates the current motion state data of the fishing boat when casting, and further combines this with theoretical analysis of the casting motion to output the current casting posture, thereby helping the user to sense whether the casting posture is correct or not, allowing the user to quickly grasp various casting postures at the early stage of learning, and reducing problems such as casting errors, unnecessary physical exhaustion, and line tangles caused by casting errors.

[0063] Those skilled in the art will understand that for convenience and brevity of explanation, the specific operation processes of the above-described systems and units may refer to the corresponding processes in the above-described method embodiments, and the description thereof will be omitted here.

[0064] The above description is merely a specific embodiment of the present invention, and the scope of protection of the present invention is not limited thereto. Those skilled in the art can easily imagine various equivalent modifications and replacements within the technical scope of the present invention, and these modifications and replacements are included in the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined only by the scope of protection of the claims.

[0065] (Addendum) (Appendix 1) training model parameters for the classification model using a historical motion state dataset of the fishing vessel at different casting attitudes; A step of acquiring current motion state data of the fishing boat during casting; inputting the current motion state data into the trained classification model to perform casting posture prediction, and outputting a probability corresponding to each casting posture; and determining the casting posture corresponding to the maximum probability as the current casting posture.

[0066] (Appendix 2) The step of acquiring current motion state data of the fishing boat during casting includes: Collecting three-axis angular acceleration data of the fishing vessel during a casting process; and extracting motion state data from the three-axis angular acceleration data.

[0067] (Appendix 3) the step of inputting the current motion state data into a trained classification model to predict a casting posture and outputting a probability corresponding to each casting posture; inputting the motion state data of the three-axis angular acceleration data into a trained classification model to predict a casting posture, and outputting a score corresponding to each casting posture; and performing probability conversion based on the score of each casting position and the total score of all casting positions, and outputting a probability corresponding to each casting position.

[0068] (Appendix 4) A step of presetting a brake mode corresponding to each casting posture; After determining that the current casting operation has ended, automatically executing a brake mode corresponding to the current casting posture, Here, the fishing state determination method described in Appendix 2 is characterized in that the method for determining the end of the current casting motion is a method for determining that the current casting motion has ended when the collected three-axis angular gravitational acceleration data switches from a moving state to a non-moving state for the first time.

[0069] (Appendix 5) collecting current acceleration data in a current casting posture, the current acceleration data including current three-axis gravitational acceleration data and current two-axis angular acceleration data, and the current two-axis angular acceleration data including current y-axis angular acceleration data and current z-axis angular acceleration data; performing a feature difference comparison between the current acceleration data and a predetermined standard acceleration data to obtain a feature score of the current acceleration data; 2. The fishing state determination method according to claim 1, further comprising the step of: providing a training suggestion based on the feature score of the current acceleration data.

[0070] (Appendix 6) the step of comparing the feature difference between the current acceleration data and preset standard acceleration data to obtain a feature score of the current acceleration data, a step of extracting current motion state data from the current acceleration data, non-motion state data within a set time length before and after the current motion state, and data at the current line winding stage in chronological order to obtain a plurality of current time series data and constructing them as a current comparison time series data set; performing at least one feature difference comparison between each current time series data in the currently compared time series data set and corresponding standard time series data in the standard acceleration data, and outputting a feature score of the corresponding item of each current time series data; and performing a weighted addition calculation on the feature scores of the same item of all the current time-series data in the currently compared dataset, and outputting a final score for each feature of the currently compared dataset.

[0071] (Appendix 7) collecting the number of casts for each casting position; When the number of times of casting of the current casting posture reaches a predetermined number, calculating an average value of each feature score in the current casting posture for the predetermined number of times; The fishing condition determination method according to claim 6, further comprising the step of determining the difference between the average value of each feature score and the corresponding standard score as the feature score compensation for the current casting posture during the subsequent casting.

[0072] (Appendix 8) Collecting a current reeling length, a current reeling speed, and a current reeling acceleration of the fishing vessel in a current casting position; Obtaining the maximum remaining unwinding length by fitting the current unwinding speed and the current unwinding acceleration according to the speed change rule; When the sum of the current unwinding length and the maximum remaining unwinding length is greater than the currently required set unwinding length, executing a brake operation corresponding to the magnitude of the brake force based on the set speed range in which the current unwinding speed is located; The fishing state determination method described in Appendix 1 further includes a step of recalculating whether the sum of the current unwinding length and the maximum remaining unwinding length exceeds the currently required set unwinding length each time the brake is operated, until the sum of the current unwinding length and the maximum remaining unwinding length approaches the currently required set unwinding length.

[0073] (Appendix 9) calculating the resting calories generated by the user throughout the fishing process; A step of obtaining the number of castings in the entire fishing process, the casting posture of each casting, and the gravity acceleration amplitude of each casting, and calculating the casting calories generated in the entire fishing process according to a preset casting calorie consumption table, wherein the casting calorie consumption table includes calories consumed by different casting postures under different gravity acceleration amplitudes; Obtaining the total reeling length of the fishing boat during the entire fishing process, and inputting the total reeling length into a preset first regression model to output the reeling calories required to reel in the total reeling length; The method further includes: obtaining the user's walking steps and walking frequency during the entire fishing process by the fishing boat; inputting the walking steps and walking frequency into a preset second regression model to output the corresponding walking calories; and calculating the sum of the resting calories, casting calories, reeling calories, and walking calories to obtain the fishing calories consumed by the user throughout the entire fishing process.

[0074] (Appendix 10) a model training unit for performing model parameter training on the classification model using a historical motion state dataset of the fishing vessel at different casting attitudes; a data acquisition unit for acquiring current motion state data of the fishing boat during casting; a posture prediction unit for inputting the current motion state data into the trained classification model to perform casting posture prediction and outputting a probability corresponding to each casting posture; a posture determining unit for determining the casting posture corresponding to the maximum probability as the current casting posture.

Claims

1. training model parameters for the classification model using a historical motion state dataset of the fishing vessel at different casting attitudes; A step of acquiring current motion state data of the fishing boat during casting; inputting the current motion state data into the trained classification model to perform casting posture prediction, and outputting a probability corresponding to each casting posture; and determining the casting posture corresponding to the maximum probability as the current casting posture.

2. The step of acquiring current motion state data of the fishing boat during casting includes: Collecting three-axis angular acceleration data of the fishing vessel during a casting process; 2. The fishing state determination method according to claim 1, further comprising: extracting motion state data from the three-axis angular acceleration data.

3. the step of inputting the current motion state data into a trained classification model to predict a casting posture and outputting a probability corresponding to each casting posture; inputting the motion state data of the three-axis angular acceleration data into a trained classification model to predict a casting posture, and outputting a score corresponding to each casting posture; and performing probability conversion based on the score of each casting position and the total score of all casting positions, and outputting a probability corresponding to each casting position.

4. A step of presetting a brake mode corresponding to each casting posture; After determining that the current casting operation has ended, automatically executing a brake mode corresponding to the current casting posture, The fishing state determination method according to claim 2, characterized in that the method for determining the end of the current casting operation is a method for determining that the current casting operation has ended when the collected three-axis angular gravitational acceleration data switches from a moving state to a non-moving state for the first time.

5. collecting current acceleration data in a current casting posture, the current acceleration data including current three-axis gravitational acceleration data and current two-axis angular acceleration data, and the current two-axis angular acceleration data including current y-axis angular acceleration data and current z-axis angular acceleration data; performing a feature difference comparison between the current acceleration data and a predetermined standard acceleration data to obtain a feature score of the current acceleration data; The fishing state determination method according to claim 1 , further comprising: providing a lesson suggestion based on the feature score of the current acceleration data.

6. the step of comparing the feature difference between the current acceleration data and preset standard acceleration data to obtain a feature score of the current acceleration data, a step of extracting current motion state data from the current acceleration data, non-motion state data within a set time length before and after the current motion state, and data at the current line winding stage in chronological order to obtain a plurality of current time series data and constructing them as a current comparison time series data set; performing at least one feature difference comparison between each current time series data in the currently compared time series data set and corresponding standard time series data in the standard acceleration data, and outputting a feature score for the corresponding item of each current time series data; and performing a weighted addition calculation on the feature scores of the same item of all the current time-series data in the currently compared data set, and outputting a final score for each feature of the currently compared data set.

7. collecting the number of casts for each casting position; When the number of times of casting of the current casting posture reaches a predetermined number, calculating an average value of each feature score in the current casting posture for the predetermined number of times; 7. The fishing condition determination method according to claim 6, further comprising the step of: determining the difference between the average value of each characteristic score and the corresponding standard score as characteristic score compensation for the current casting posture during a subsequent casting.

8. Collecting a current reeling length, a current reeling speed, and a current reeling acceleration of the fishing vessel in a current casting position; Obtaining the maximum remaining unwinding length by fitting the current unwinding speed and the current unwinding acceleration according to the speed change rule; When the sum of the current unwinding length and the maximum remaining unwinding length is greater than the currently required set unwinding length, executing a brake operation corresponding to the magnitude of the brake force based on the set speed range in which the current unwinding speed is located; The fishing state determination method according to claim 1, further comprising a step of recalculating whether the sum of the current unwinding length and the maximum remaining unwinding length exceeds the currently required set unwinding length each time the brake is operated, until the sum of the current unwinding length and the maximum remaining unwinding length approaches the currently required set unwinding length.

9. calculating the resting calories generated by the user throughout the fishing process; A step of obtaining the number of castings in the entire fishing process, the casting posture of each casting, and the gravity acceleration amplitude of each casting, and calculating the casting calories generated in the entire fishing process according to a preset casting calorie consumption table, wherein the casting calorie consumption table includes calories consumed by different casting postures under different gravity acceleration amplitudes; Obtaining the total reeling length of the fishing boat during the entire fishing process, and inputting the total reeling length into a preset first regression model to output the reeling calories required to reel in the total reeling length; The method further includes: obtaining the user's walking steps and walking frequency during the entire fishing process by the fishing boat; inputting the walking steps and walking frequency into a second regression model that is preset, and outputting the corresponding walking calories; and calculating the sum of the resting calories, casting calories, reeling calories and walking calories to obtain the fishing calories consumed by the user throughout the entire fishing process.

10. a model training unit for performing model parameter training on the classification model using a historical motion state dataset of the fishing vessel at different casting attitudes; a data acquisition unit for acquiring current motion state data of the fishing boat during casting; a posture prediction unit for inputting the current motion state data into the trained classification model to perform casting posture prediction and outputting a probability corresponding to each casting posture; a posture determining unit for determining the casting posture corresponding to the maximum probability as the current casting posture.

Images