Method and system for monitoring ingestion of shrimps and crabs based on passive acoustic technology

Abstract

The invention discloses a shrimp and crab feeding monitoring method and system based on a passive acoustic technology, and belongs to the technical field of aquaculture. According to the method, breeding acoustic signals are collected, shrimp and crab feeding acoustic signals are extracted through preprocessing, and then a feeding time-frequency distribution diagram is obtained through time-frequency analysis. A standard time-frequency distribution map library containing different feeding grades is constructed, an image hash technology is used for processing actual feeding signals and signals in the standard map library, fingerprints are obtained and matched, and therefore the feeding intensity of shrimps and crabs is accurately determined. According to the invention, the feeding activity of prawns and crabs can be accurately and effectively monitored, and the feeding strategy can be automatically adjusted according to the feeding intensity.

Description

Technical Field

[0001] The present invention relates to the technical field of aquaculture, and more particularly to a shrimp and crab feeding monitoring method and system based on passive acoustic technology. Background Art

[0002] In aquaculture, accurate monitoring of shrimp and crab feeding behavior is of far-reaching and wide-ranging importance for optimizing feed feeding strategies, improving aquaculture efficiency and yield. This is not only related to the direct economic benefits of aquaculture, but also to the efficient use of resources, sustainable protection of the environment and the healthy development of the entire aquaculture industry chain.

[0003] At present, traditional methods face many challenges when monitoring the feeding behavior of shrimps and crabs in aquaculture, including the subjectivity and limitations of manual observation leading to inaccurate judgments, complex physical monitoring methods and susceptibility to environmental interference, low accuracy and reliability of monitoring data, and lack of intelligent and automated means for automatic identification and classification. These problems together limit farmers' in-depth understanding and optimized management of shrimp and crab feeding behavior.

[0004] Therefore, how to propose a shrimp and crab feeding monitoring method and system based on passive acoustic technology to achieve accurate monitoring of shrimp and crab feeding behavior is a problem that technical personnel in this field urgently need to solve. Summary of the invention

[0005] In view of this, the present invention provides a shrimp and crab feeding monitoring method and system based on passive acoustic technology to improve the accuracy of shrimp and crab feeding monitoring.

[0006] In order to achieve the above object, the present invention adopts the following technical solution:

[0007] On the one hand, the present invention provides a shrimp and crab feeding monitoring method based on passive acoustic technology, comprising the following steps:

[0008] Collecting aquaculture acoustic signals, preprocessing the aquaculture acoustic signals, and obtaining shrimp and crab feeding sound signals;

[0009] Performing time-frequency analysis on the shrimp and crab feeding sound signals to obtain a feeding time-frequency distribution diagram;

[0010] Constructing a food intake standard library, wherein the food intake standard library includes standard time-frequency distribution diagrams of different food intake levels;

[0011] Processing the time-frequency distribution graph based on image hashing technology to obtain a fingerprint of the time-frequency distribution graph; the time-frequency distribution graph includes a feeding time-frequency distribution graph and standard time-frequency distribution graphs of different feeding levels;

[0012] Matching the fingerprints of the time-frequency distribution graph, and determining the feeding intensity according to the matching result;

[0013] The shrimps and crabs are automatically fed according to the feeding intensity.

[0014] Preferably, the aquaculture acoustic signal is preprocessed to obtain shrimp and crab feeding sound signals, including:

[0015] Performing denoising and filtering processing on the aquaculture acoustic signal to obtain an acoustic signal to be processed;

[0016] Extracting time-frequency features of the sound signal to be processed;

[0017] Inputting the time-frequency features into a trained deep learning-based recognition model to obtain a recognition result;

[0018] According to the recognition result, the shrimp and crab feeding sound signal and the shrimp and crab category corresponding to the shrimp and crab feeding sound signal are obtained.

[0019] Preferably, the time-frequency distribution graph is processed based on the image hashing technology to obtain the fingerprint of the time-frequency distribution graph, including:

[0020] The time-frequency distribution graph is compressed into a small picture of n×n size, the small picture is converted into a grayscale image, and the matrix I X express;

[0021] Calculate the matrix I X The element average value of , compare the grayscale of each pixel with the element average value to generate a new matrix T X ;

[0022] The new matrix T X Combine them from top to bottom and from left to right to get n 2 A bit integer sequence is used as the fingerprint string of the time-frequency distribution graph.

[0023] Preferably, the element average value u x The calculation formula is as follows:

[0024]

[0025] In the formula, I X (i,j) is the matrix element value, i∈[1,n], j∈[1,n].

[0026] Preferably, the new matrix T X The generation formula is as follows:

[0027]

[0028] Where, T X (i,j) is the new matrix element value.

[0029] Preferably, matching the fingerprints of the time-frequency distribution graph and determining the feeding intensity according to the matching results comprises:

[0030] The fingerprint string of the feeding time-frequency distribution diagram is compared with the fingerprint strings of the standard time-frequency distribution diagrams of different feeding levels, and the fingerprint strings of the standard time-frequency distribution diagrams of different feeding levels and the fingerprint strings of the feeding time-frequency distribution diagrams have the same value at the same position, and they are sorted from large to small. The feeding level corresponding to the fingerprint string of the standard time-frequency distribution diagram that is ranked first is the current feeding intensity of shrimps and crabs.

[0031] Preferably, automatically feeding shrimps and crabs according to the feeding intensity comprises:

[0032] Establish the corresponding relationship between feeding level and feeding amount;

[0033] The feed feeding amount is obtained according to the corresponding relationship and the feeding level corresponding to the feeding intensity, and automatic feeding is performed.

[0034] On the other hand, the present invention also proposes a shrimp and crab feeding monitoring system based on passive acoustic technology, which is used to implement the above shrimp and crab feeding monitoring method, comprising:

[0035] The acoustic signal acquisition module collects the aquaculture acoustic signals and pre-processes the aquaculture acoustic signals to obtain the shrimp and crab feeding acoustic signals;

[0036] A time-frequency analysis module performs time-frequency analysis on the shrimp and crab feeding sound signals to obtain a feeding time-frequency distribution diagram;

[0037] A standard picture library construction module is used to construct a feeding standard picture library, wherein the feeding standard picture library includes standard time-frequency distribution diagrams of different feeding levels;

[0038] A fingerprint acquisition module processes the time-frequency distribution graph based on the image hashing technology to obtain the fingerprint of the time-frequency distribution graph; the time-frequency distribution graph includes a feeding time-frequency distribution graph and a standard time-frequency distribution graph of different feeding levels;

[0039] A matching module, matching the fingerprint of the time-frequency distribution graph, and determining the feeding intensity according to the matching result;

[0040] The automatic feeding module automatically feeds shrimps and crabs according to the feeding intensity.

[0041] Preferably, the acoustic signal acquisition module includes:

[0042] A preprocessing unit performs denoising and filtering on the aquaculture acoustic signal to obtain an acoustic signal to be processed;

[0043] A feature extraction unit, extracting time-frequency features of the sound signal to be processed;

[0044] A recognition unit, inputting the time-frequency features into a trained recognition model based on deep learning to obtain a recognition result;

[0045] The sound signal acquisition unit obtains the shrimp and crab feeding sound signal and the shrimp and crab category corresponding to the shrimp and crab feeding sound signal according to the recognition result.

[0046] Preferably, the fingerprint acquisition module includes:

[0047] The image compression unit compresses the time-frequency distribution diagram into a small picture of n×n size, converts the small picture into a grayscale image, and converts the grayscale image into a grayscale image in the form of a matrix I X express;

[0048] Matrix transformation unit, calculates the matrix I X The element average value of , compare the grayscale of each pixel with the element average value to generate a new matrix T X ;

[0049] The fingerprint generation unit converts the new matrix T X Combine them from top to bottom and from left to right to get n 2 A bit integer sequence is used as the fingerprint string of the time-frequency distribution graph.

[0050] It can be seen from the above technical solutions that, compared with the prior art, the present invention discloses a shrimp and crab feeding monitoring method and system based on passive acoustic technology, which improves the accuracy and efficiency of shrimp and crab feeding monitoring by utilizing acoustic signals and time-frequency analysis technology; by constructing a standard gallery and fingerprint matching, a quantitative evaluation of feeding intensity is achieved; the formulation of automated feeding strategies helps to optimize breeding management and reduce resource waste, while improving the growth performance and breeding benefits of shrimp and crabs, providing strong support for the refined management of shrimp and crab breeding. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.

[0052] Figure 1 A flow chart of the method provided by the present invention;

[0053] Figure 2 This is a system architecture diagram provided by the present invention. DETAILED DESCRIPTION

[0054] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0055] On the one hand, the present invention proposes a shrimp and crab feeding monitoring method based on passive acoustic technology, referring to Figure 1 , the method comprises the following steps:

[0056] S1. Collect aquaculture acoustic signals, pre-process the aquaculture acoustic signals, and obtain shrimp and crab feeding sound signals, including:

[0057] S11. De-noise and filter the aquaculture acoustic signal to obtain the acoustic signal to be processed.

[0058] First, analyze the background noise characteristics in the aquaculture environment and identify the main noise sources (such as water flow noise, mechanical noise, etc.); then use adaptive filters, spectral subtraction or other noise reduction algorithms (such as wavelet denoising) to remove the background noise. Adaptive filters can adjust parameters in real time to adapt to the changing noise environment.

[0059] Determine the frequency range of shrimp and crab feeding sound signals, and select a suitable bandpass filter to only allow signals within this frequency range to pass, thereby further removing interference signals of other frequencies.

[0060] S12. Extracting time-frequency features of the sound signal to be processed.

[0061] Features extracted from the time-frequency analysis results may include energy distribution, frequency center of gravity, spectral entropy, etc. of the time-frequency graph. Energy distribution can reflect the energy of signals in different frequency and time regions, which is helpful for identifying energy concentration areas in feeding behavior (such as high-intensity signal periods generated by feeding actions). Frequency center of gravity can describe the average position of signal frequencies, which is meaningful for determining the main frequency components of feeding sounds. Spectral entropy can measure the complexity of the signal spectrum, and different feeding behaviors may correspond to different spectral entropy values. These features can be combined into a feature vector for subsequent model input.

[0062] S13. Input the time-frequency features into the trained deep learning-based recognition model to obtain the recognition results.

[0063] The recognition model can select deep learning models such as convolutional neural network (CNN), recurrent neural network (RNN) and its variants (such as long short-term memory network LSTM, gated recurrent unit GRU) and so on.

[0064] Before training the recognition model, a large amount of shrimp and crab feeding sound signal data needs to be collected and annotated, including the category of feeding sound signals (feeding sounds of different types of shrimp and crabs), feeding status (such as starting feeding, continuing feeding, stopping feeding, etc.) and other information. The data is divided into training set, validation set and test set, and the model is trained using the training set. The weight and bias of the model are adjusted through the back propagation algorithm to minimize the error between the model's prediction results and the true label. During the training process, a variety of optimization algorithms can be used, such as stochastic gradient descent (SGD), Adagrad, Adadelta, Adam, etc., and appropriate hyperparameters such as learning rate and momentum can be selected to accelerate the model convergence speed and improve model performance. At the same time, regularization technology can also be used to prevent model overfitting and improve the generalization ability of the model. By evaluating the performance of the model on the validation set, the model parameters are adjusted according to indicators such as accuracy, recall rate, F1 value, etc. When the performance of the model on the validation set no longer improves, the training is stopped, and the model with the best performance is selected for test set evaluation and practical application.

[0065] S14. Obtain the shrimp and crab feeding sound signals and the shrimp and crab categories corresponding to the shrimp and crab feeding sound signals according to the recognition results.

[0066] S2. Perform time-frequency analysis on the feeding sound signals of shrimps and crabs to obtain the feeding time-frequency distribution diagram.

[0067] For the collected shrimp and crab sound signals, its short-time Fourier transform is defined as:

[0068]

[0069] Among them, h(t) is the short-time Fourier transform window function; τ represents the signal in the τ period around the original signal at time t. The window function h(t) continuously translates on the time axis with time t, and the time domain signal (i.e., shrimp and crab sound signal) s(t) of the corresponding time period continuously enters the state of being analyzed, and finally reflects the time-frequency distribution of the original signal. By calculating STFT s (t,ω), we can get the energy distribution of shrimp and crab sound signals at different times t and frequencies ω, and thus obtain the time-frequency distribution diagram.

[0070] In the selection of time-frequency analysis parameters (such as window function type, window length, etc.), it is necessary to select according to the frequency range and variation characteristics of shrimp and crab sound signals to better capture the time-frequency characteristics related to feeding behavior. For example, if the frequency of shrimp and crab feeding sounds is mainly concentrated in the lower frequency band, a window function with higher resolution for low-frequency signals can be selected.

[0071] In the monitoring of shrimp and crab feeding, it is necessary to pay attention to the performance of the sound characteristics unique to the feeding behavior in the time-frequency diagram, such as the specific frequency pulse signal generated by the feeding action and its distribution pattern in the time-frequency diagram.

[0072] S3. Construct a standard feeding picture library, which includes standard time-frequency distribution diagrams of different feeding levels.

[0073] The acoustic signals of shrimps and crabs in different feeding scenarios were extensively collected, and the corresponding time-frequency distribution diagram was obtained through the above-mentioned time-frequency analysis.

[0074] These time-frequency distribution diagrams were screened and sorted to establish a standard time-frequency distribution diagram library containing various feeding states and shrimp and crab species as a benchmark for subsequent matching.

[0075] S4. Process the time-frequency distribution graph based on the image hashing technology to obtain the fingerprint of the time-frequency distribution graph. The time-frequency distribution graph includes the feeding time-frequency distribution graph and the standard time-frequency distribution graphs of different feeding levels.

[0076] This step includes:

[0077] S41. Compress the time-frequency distribution graph into a small picture of n×n size. The smaller the size, the less image details are retained and the less clear the image is. There are a total of n 2 pixels.

[0078] Convert the n×n small image to a grayscale image and use the matrix I X express.

[0079] S42. Calculate matrix I X The element average value of , compare the grayscale of each pixel with the element average value to generate a new matrix T X .

[0080] Among them, the element average value u x The calculation formula is as follows:

[0081]

[0082] In the formula, I X (i,j) is the matrix element value, i∈[1,n], j∈[1,n].

[0083] The new matrix T X The generation formula is as follows:

[0084]

[0085] Where, T X (i,j) is the new matrix element value.

[0086] S43. The new matrix T X According to the principle of combining from top to bottom (i=i→n) and from left to right (j=i→n), we get n 2 The bit integer sequence is used as the fingerprint string of the time-frequency distribution graph.

[0087] S5. Match the fingerprints of the time-frequency distribution diagram and determine the feeding intensity according to the matching results.

[0088] The fingerprint string of the feeding time-frequency distribution diagram is compared with the fingerprint strings of the standard time-frequency distribution diagrams of different feeding levels, and the fingerprint strings of the standard time-frequency distribution diagrams of different feeding levels are obtained. The values ​​of the fingerprint strings of the feeding time-frequency distribution diagrams at the same position are the same, and they are sorted from large to small. The feeding level corresponding to the fingerprint string of the standard time-frequency distribution diagram that is ranked first is the current feeding intensity of shrimps and crabs.

[0089] S6. Automatically feed shrimps and crabs according to feeding intensity, including:

[0090] S61. Construct a corresponding relationship between feeding level and feeding amount.

[0091] Collect actual food intake data at various feeding intensities and record feed consumption at different feeding levels; conduct multiple experiments to ensure the reliability of the data.

[0092] In this embodiment, linear regression, curve fitting or machine learning models (such as decision trees or neural networks) can be used to establish the relationship between feeding intensity and feeding amount.

[0093] Furthermore, since the breeding environment and the growth status of shrimps and crabs are changing dynamically, a dynamic adjustment mechanism needs to be established to update the corresponding relationship between feeding level and feeding amount. The growth status of shrimps and crabs and changes in environmental conditions should be reassessed regularly, and the model should be revised using new data.

[0094] S62. Obtain the feed amount according to the corresponding relationship and the feeding level corresponding to the feeding intensity, and perform automatic feeding.

[0095] On the other hand, the present invention also proposes a shrimp and crab feeding monitoring system based on passive acoustic technology, which is used to implement the above shrimp and crab feeding monitoring method. Figure 2 As shown, the system includes:

[0096] The acoustic signal acquisition module collects the aquaculture acoustic signals, pre-processes the aquaculture acoustic signals, and obtains the shrimp and crab feeding acoustic signals;

[0097] The time-frequency analysis module performs time-frequency analysis on the feeding sound signals of shrimps and crabs to obtain the feeding time-frequency distribution diagram;

[0098] A standard image library construction module is used to construct a feeding standard image library, which includes standard time-frequency distribution diagrams of different feeding levels;

[0099] A fingerprint acquisition module processes the time-frequency distribution graph based on the image hashing technology to obtain the fingerprint of the time-frequency distribution graph; the time-frequency distribution graph includes a feeding time-frequency distribution graph and a standard time-frequency distribution graph of different feeding levels;

[0100] A matching module matches the fingerprint of the time-frequency distribution graph and determines the feeding intensity according to the matching result;

[0101] Automatic feeding module automatically feeds shrimps and crabs according to feeding intensity.

[0102] Preferably, the acoustic signal acquisition module includes:

[0103] The preprocessing unit performs denoising and filtering on the aquaculture acoustic signal to obtain the acoustic signal to be processed;

[0104] A feature extraction unit, which extracts the time-frequency features of the sound signal to be processed;

[0105] The recognition unit inputs the time-frequency features into the trained deep learning-based recognition model to obtain the recognition results;

[0106] The sound signal acquisition unit obtains the shrimp and crab feeding sound signals and the shrimp and crab categories corresponding to the shrimp and crab feeding sound signals according to the recognition results.

[0107] Preferably, the fingerprint acquisition module includes:

[0108] The image compression unit compresses the time-frequency distribution map into a small image of size n×n, converts the small image into a grayscale image, and converts it into a grayscale image in the matrix I X express;

[0109] Matrix transformation unit, calculates matrix I X The element average value of , compare the grayscale of each pixel with the element average value to generate a new matrix T X ;

[0110] The fingerprint generation unit converts the new matrix T X Combine them from top to bottom and from left to right to get n 2 The bit integer sequence is used as the fingerprint string of the time-frequency distribution graph.

[0111] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.

[0112] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A shrimp and crab feeding monitoring method based on passive acoustic technology, characterized in that: The following steps are involved: Collecting aquaculture acoustic signals, preprocessing the aquaculture acoustic signals, and obtaining shrimp and crab feeding sound signals; Performing time-frequency analysis on the shrimp and crab feeding sound signals to obtain a feeding time-frequency distribution diagram; Constructing a food intake standard library, wherein the food intake standard library includes standard time-frequency distribution diagrams of different food intake levels; Processing the time-frequency distribution graph based on image hashing technology to obtain a fingerprint of the time-frequency distribution graph; the time-frequency distribution graph includes a feeding time-frequency distribution graph and standard time-frequency distribution graphs of different feeding levels; Matching the fingerprints of the time-frequency distribution graph, and determining the feeding intensity according to the matching result; The shrimps and crabs are automatically fed according to the feeding intensity.

2. The method for monitoring shrimp and crab feeding based on passive acoustic technology according to claim 1 is characterized in that: Preprocessing the aquaculture acoustic signal to obtain shrimp and crab feeding sound signals includes: De-noising and filtering the aquaculture acoustic signal to obtain an acoustic signal to be processed; Extracting time-frequency features of the sound signal to be processed; Inputting the time-frequency features into a trained deep learning-based recognition model to obtain a recognition result; The shrimp and crab feeding sound signals and the shrimp and crab categories corresponding to the shrimp and crab feeding sound signals are obtained according to the recognition results.

3. The shrimp and crab feeding monitoring method based on passive acoustic technology according to claim 1 is characterized in that: The time-frequency distribution graph is processed based on the image hashing technology to obtain the fingerprint of the time-frequency distribution graph, including: The time-frequency distribution graph is compressed into a small picture of n×n size, the small picture is converted into a grayscale image, and the matrix I X express; Calculate the matrix I X The element average value of , compare the grayscale of each pixel with the element average value to generate a new matrix T X ; The new matrix T X Combine them from top to bottom and from left to right to get n 2 A bit integer sequence is used as the fingerprint string of the time-frequency distribution graph.

4. The shrimp and crab feeding monitoring method based on passive acoustic technology according to claim 3 is characterized in that: The element average value u x The calculation formula is as follows: In the formula, I X (i,j) is the matrix element value, i∈[1,n], j∈[1,n].

5. The shrimp and crab feeding monitoring method based on passive acoustic technology according to claim 4 is characterized in that: The new matrix T X The generation formula is as follows: Where, T X (i,j) is the new matrix element value.

6. The shrimp and crab feeding monitoring method based on passive acoustic technology according to claim 3 is characterized in that: Matching the fingerprint of the time-frequency distribution graph and determining the feeding intensity according to the matching result includes: The fingerprint string of the feeding time-frequency distribution diagram is compared with the fingerprint strings of the standard time-frequency distribution diagrams of different feeding levels, and the fingerprint strings of the standard time-frequency distribution diagrams of different feeding levels and the fingerprint strings of the feeding time-frequency distribution diagrams have the same value at the same position, and they are sorted from large to small. The feeding level corresponding to the fingerprint string of the standard time-frequency distribution diagram that is ranked first is the current feeding intensity of shrimps and crabs.

7. The shrimp and crab feeding monitoring method based on passive acoustic technology according to claim 3 is characterized in that: Automatically feeding shrimps and crabs according to the feeding intensity includes: Establish the corresponding relationship between feeding level and feeding amount; The feed feeding amount is obtained according to the corresponding relationship and the feeding level corresponding to the feeding intensity, and automatic feeding is performed.

8. A shrimp and crab feeding monitoring system based on passive acoustic technology, characterized in that: include: The acoustic signal acquisition module collects the aquaculture acoustic signals and pre-processes the aquaculture acoustic signals to obtain the shrimp and crab feeding acoustic signals; A time-frequency analysis module performs time-frequency analysis on the shrimp and crab feeding sound signals to obtain a feeding time-frequency distribution diagram; A standard picture library construction module is used to construct a feeding standard picture library, wherein the feeding standard picture library includes standard time-frequency distribution diagrams of different feeding levels; A fingerprint acquisition module processes the time-frequency distribution graph based on the image hashing technology to obtain the fingerprint of the time-frequency distribution graph; the time-frequency distribution graph includes a feeding time-frequency distribution graph and a standard time-frequency distribution graph of different feeding levels; A matching module, matching the fingerprint of the time-frequency distribution graph, and determining the feeding intensity according to the matching result; The automatic feeding module automatically feeds shrimps and crabs according to the feeding intensity.

9. The method for monitoring shrimp and crab feeding based on passive acoustic technology according to claim 8, characterized in that: The acoustic signal acquisition module comprises: A preprocessing unit performs denoising and filtering on the aquaculture acoustic signal to obtain an acoustic signal to be processed; A feature extraction unit, extracting time-frequency features of the sound signal to be processed; A recognition unit, inputting the time-frequency features into a trained recognition model based on deep learning to obtain a recognition result; The sound signal acquisition unit obtains the shrimp and crab feeding sound signal and the shrimp and crab category corresponding to the shrimp and crab feeding sound signal according to the recognition result.

10. The shrimp and crab feeding monitoring method based on passive acoustic technology according to claim 8, characterized in that: The fingerprint acquisition module comprises: The image compression unit compresses the time-frequency distribution diagram into a small picture of n×n size, converts the small picture into a grayscale image, and converts the grayscale image into a grayscale image in the form of a matrix I X express; Matrix transformation unit, calculates the matrix I X The element average value of , compare the grayscale of each pixel with the element average value to generate a new matrix T X ; The fingerprint generation unit converts the new matrix T X Combine them from top to bottom and from left to right to get n 2 A bit integer sequence is used as the fingerprint string of the time-frequency distribution graph.

Images