CN117611983B - Underwater target detection method and system based on hidden communication technology and deep learning - Google Patents

Abstract

The invention discloses an underwater target detection method and system based on a hidden communication technology and deep learning, and the method comprises the following steps: acquiring an image of a target object in real time; preprocessing the image, and extracting features of the preprocessed image by using a convolutional neural network model to obtain boundary frame coordinates, category labels and confidence scores of each target; the information transmission method based on the hidden communication realizes the safe transmission of the boundary frame coordinates, the category labels, the confidence scores and the communication information of each target. The invention can effectively solve the problems of real-time image processing and underwater hidden communication of the underwater robot, and is more in line with the complex and changeable environment at present.

Description

Underwater target detection method and system based on covert communication technology and deep learning

Technical Field

The present invention belongs to the field of underwater robots, and in particular to the fields of underwater target detection and covert communication, and specifically to an underwater target detection method and system based on covert communication technology and deep learning.

Background technique

With the development of science and technology, underwater target detection has become a hot technical issue. Due to the particularity of the underwater environment, underwater image acquisition is much more difficult than land images. The acquisition of underwater images is affected by many factors, such as the complex underwater lighting environment, water quality conditions, the similarity between the target and the background, etc. These are the main factors affecting the accuracy of underwater target detection. In recent years, deep learning is booming. Target detection technology has made great progress and has been widely used in various scenarios. However, the current method has a complex network structure and a large number of parameters, which is not conducive to real-time detection.

Underwater communication technology refers to the technology of transmitting information in an underwater environment, including acoustic communication, electromagnetic wave communication, optical communication and other forms. In seawater, due to the serious absorption and attenuation of electromagnetic waves and light waves in water, sound waves have become the main way for information to be transmitted in water. In underwater environments, traditional communication methods such as electromagnetic waves and light waves have many limitations and shortcomings and cannot meet the needs of actual applications. However, underwater acoustic covert communication technology can effectively improve the confidentiality and concealment of underwater communications by using special modulation and demodulation methods, encryption algorithms and other means, and has stronger anti-interference capabilities. It can cope with factors such as ocean background noise, interference from underwater biological activities, and hostile interference to ensure accurate transmission of information.

The coding layer covert communication technology transmits secret information by embedding it into the coded public information to achieve covert communication. The embedding of secret information will affect the error correction performance of channel coding, thereby increasing the bit error rate of the information transmitted in the original communication system at the receiving end and the risk of the embedded information being detected. In order to reduce the negative impact of the embedding of secret information on the original communication system, a binary Hamming matrix embedding method is introduced into the covert communication of the coding layer. This embedding method can effectively reduce the impact of secret information on the carrier information and improve the concealment and security of covert communication. However, the introduction of the binary Hamming matrix embedding method will increase the bit error rate of secret information.

Therefore, the present invention proposes an underwater robot solution combining a technology for real-time underwater target detection and an improved coding layer covert communication technology based on binary Hamming matrix embedding for underwater acoustic communication, which can solve this problem well.

Summary of the invention

In view of the shortcomings of the existing technology, the present invention proposes an underwater target detection method and system based on covert communication technology and deep learning, which can effectively solve the real-time image processing problems of underwater robots and underwater covert communication problems, and is more in line with the current complex and changeable environment.

To achieve the above object, the present invention provides the following solutions:

The underwater target detection method based on covert communication technology and deep learning includes the following steps:

Collect images of target objects in real time;

Preprocessing the image, and extracting features from the preprocessed image using a convolutional neural network model to obtain bounding box coordinates, category labels, and confidence scores for each target;

The information transmission method based on covert communication realizes the secure transmission of bounding box coordinates, category label and confidence score of each target and communication information.

Preferably, the method for collecting an image of a target object in real time includes:

Use a high-definition camera module to capture images of the area to be inspected and obtain images of the target object.

Preferably, the method of preprocessing the image and extracting features from the preprocessed image using a convolutional neural network model to obtain the bounding box coordinates, category label and confidence score of each target includes:

Adjust the size of the image of the target object and normalize the pixel values;

Based on the YOLOv8 model, multiple fine-grained feature layers are introduced on the basis of Darknet53. The feature information of different scales of the preprocessed image is extracted and fused. Multiple detection heads are used to predict target boxes of different scales to obtain the bounding box coordinates, category label and confidence score of each target.

Preferably, the method for realizing the secure transmission of the bounding box coordinates, category label and confidence score of each target and the communication information based on the covert communication information transmission method includes:

Perform convolution coding on public information to form carrier information; perform convolution coding on secret information and interleave the secret information through an interleaver; embed the secret information into public information using a matrix embedding method based on binary Hamming code to obtain carrier information; perform signal modulation on the carrier information to obtain a modulated signal and send it out through sonar; wherein the secret information is information processed by an image processing module;

The secret information is extracted by matrix quantization, and the result of the matrix quantization extraction is deinterleaved as the input of the Viterbi soft decision decoder to decode and recover the secret information.

The present invention also discloses an underwater target detection system based on covert communication technology and deep learning, comprising: an image processing module, an image acquisition module and a covert communication module;

The image acquisition module is used to acquire images of target objects in real time;

The image processing module is used to preprocess the image, and use a convolutional neural network model to extract features from the preprocessed image to obtain the bounding box coordinates, category label and confidence score of each target;

The covert communication module is used to achieve the secure transmission of the bounding box coordinates, category label and confidence score of each target and the communication information based on the covert communication information transmission method.

Preferably, the image acquisition module includes: a high-definition camera module and an LED fill light device;

The high-definition camera module is used to collect images of the area to be detected and obtain images of the target object;

The LED fill light device is used to assist the high-definition camera module.

Preferably, the image processing module includes: a preprocessing unit and a feature extraction unit;

The preprocessing unit is used to adjust the size of the image of the target object and normalize the pixel values;

The feature extraction unit is used to introduce multiple fine-grained feature layers based on the YOLOv8 model on the basis of Darknet53, extract feature information of different scales of the preprocessed image and fuse them, use multiple detection heads to predict target frames of different scales, and obtain the bounding box coordinates, category label and confidence score of each target.

Preferably, the covert communication module comprises: a transmitter, a receiver, a sonar, a signal processing unit and a control unit;

The signal processing unit is used to perform convolution coding operation on public information to form carrier information; perform convolution coding operation on secret information and interleave the secret information through an interleaver; embed the secret information into the public information using a matrix embedding method based on binary Hamming code to obtain carrier information; perform signal modulation on the carrier information to obtain a modulated signal; wherein the secret information is information processed by the image processing module;

The transmitter is used to send the secret information;

The sonar is used to send out the modulated signal;

The receiver is used to receive the data signal from the transmitter;

The control unit is used to perform matrix quantization extraction on the secret information, de-interleave the result of the matrix quantization extraction as the input of the Viterbi soft decision decoder, and decode and restore the secret information.

The present invention also discloses an underwater target detection robot based on covert communication technology and deep learning, and the robot applies the underwater target detection method based on covert communication technology and deep learning.

Compared with the prior art, the present invention has the following beneficial effects:

The image processing module of the present invention is based on the YOLOv8 algorithm, adopts DarkNet53 as the feature extractor, and introduces the strategies of multi-scale prediction and feature fusion, which can better capture the target information of different scales. By performing target detection at multiple different scales, it can effectively process targets of different sizes, perform detection at different levels of the network at the same time, and improve the accuracy and stability of the detection results through feature fusion.

The communication module of the present invention adopts the covert communication technology based on the coding layer, and introduces the binary Hamming matrix embedding method for reducing the modification amount into the covert communication system of the coding layer. This embedding method can reduce the modification amount of information embedding to the carrier information. At the same time, by introducing the interleaving module, the extraction scheme is improved to reduce the negative effect of this embedding method on the reliability of secret information. The role of interleaving is to convert the continuous errors caused by this embedding method into random errors, which is beneficial to the decoding of convolutional codes and increases the reliability of secret information. The matrix quantization extraction is an extraction scheme based on this embedding method that enables the extracted information to be soft-decision decoded, which can further reduce the bit error rate of secret information.

BRIEF DESCRIPTION OF THE DRAWINGS

In order to more clearly illustrate the technical solution of the present invention, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

FIG1 is an overall framework diagram of an underwater target detection system based on covert communication technology and deep learning in an embodiment of the present invention;

FIG2 is an overall framework diagram of an image processing module network model in an embodiment of the present invention;

FIG3 is a structural diagram of an image acquisition module in an embodiment of the present invention;

FIG4 is a system block diagram of a coding layer covert communication solution in an embodiment of the present invention.

FIG. 5 is a schematic diagram of a mechanical housing of an underwater target detection robot in an embodiment of the present invention.

Detailed ways

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.

In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

Embodiment 1

The present invention provides an underwater target detection method based on covert communication technology and deep learning, comprising the following steps:

Collect images of target objects in real time;

Preprocess the image and use the convolutional neural network model to extract features from the preprocessed image to obtain the bounding box coordinates, category label, and confidence score of each object;

The information transmission method based on covert communication realizes the secure transmission of bounding box coordinates, category label and confidence score of each target and communication information.

In this embodiment, according to the requirements of the target detection task, image data containing the target object is collected. The data can be obtained from a public data set, or collected or purchased by yourself. Ensure that the data set contains diversity and representativeness of each category of targets.

Use LabelImg or RectLabel to annotate the collected images and mark the location and category of each target. Common target annotation formats include Bounding Box and Mask.

Divide the labeled data set into training set, validation set and test set. Usually use 70% to 80% of the data as training set, 10% to 15% of the data as validation set, and the remaining data as test set. Ensure that the target categories of each data set are relatively evenly distributed.

Apply data augmentation techniques to the training set to increase the diversity and quantity of the data set and improve the generalization ability of the model. Common data augmentation methods include random cropping, scaling, translation, rotation, brightness adjustment, color transformation, etc.

Convert annotation and image data into the format required for model training. For the YOLOv8 model, the commonly used data format is Darknet format or COCO format. You can use the corresponding tools to convert the dataset to the specified format.

During the training process, the image data is preprocessed, such as normalization, scaling to a fixed size, etc. This can speed up the training process and improve the performance and stability of the model.

Use the data loading tool DataLoader provided by the deep learning framework PyTorch to load the prepared dataset into the model for training.

The above is the data collection and preprocessing process, followed by the training process of a specific target detection model.

The network structure of the YOLOv8 model is built using the deep learning framework PyTorch. The construction process includes the combination of downsampling network, feature extraction layer and detection head.

Initialize the model weights randomly or with pre-trained weights.

Define the loss function of the YOLOv8 model. The loss function of YOLOv8 usually consists of position loss, classification loss, and confidence loss. You can use existing loss functions, such as cross entropy, and adjust them accordingly according to task requirements.

Use the data loading tool PyTorch's DataLoader to load the prepared training set and validation set data.

Select an appropriate optimizer, here we choose the SGD optimizer, and then set the corresponding hyperparameters such as learning rate, weight decay, momentum, etc. These hyperparameters can be debugged and optimized through experiments.

In each training cycle, a batch of training data is input into the model for forward propagation calculation, and then backpropagation is performed to update the model parameters. The model weights are updated according to the value of the loss function.

The rate of model training can be controlled by learning rate decay, which usually gradually reduces the learning rate during training.

Regularly save the model weights during the training process for subsequent model evaluation and inference.

Deploy the saved model to the operating system of the underwater robot for use in underwater target detection.

Underwater target detection requires images collected by underwater robots in real time. First, the images collected by underwater robots are preprocessed, including image size adjustment, pixel value normalization, etc. These preprocessing operations help improve the performance and accuracy of the model.

The convolutional neural network (CNN) model is used to extract features from the preprocessed images. YOLOv8 uses Darknet53 as the feature extraction network, which consists of multiple convolutional layers, residual blocks, and pooling layers.

YOLOv8 introduces multiple fine-grained features based on Darknet53 to extract feature information of different scales. By fusing features at different levels, the contextual information of objects of different sizes can be better captured.

YOLOv8 uses multiple detection heads to predict object boxes of different scales. Each detection head is responsible for predicting objects of a specific scale and outputs the corresponding classification probability, bounding box location, and confidence score.

The predictions output by the detection head are decoded to obtain the bounding box coordinates, category label, and confidence score of each object.

Since the same object may be detected by multiple bounding boxes, in order to eliminate duplicate detection results, the NMS algorithm is used to filter the predicted boxes. NMS selects the most likely target box based on the confidence score and overlap of the predicted boxes.

The target frame after NMS processing is further processed, such as setting a confidence threshold to filter out low-confidence target frames.

Finally, the output includes the category label, location information, and confidence score of the detected target box.

In this embodiment, the device for collecting images of target objects in real time mainly includes two major components: a high-definition camera module and an LED fill light device.

The underwater robot is remotely controlled to collect images of the inspection area and temporarily store the collected images in the storage system of the underwater robot.

In this embodiment, the information transmission method based on covert communication realizes the secure transmission of the bounding box coordinates, category label and confidence score of each target and the communication information, which mainly includes the following components: a transmitter, a receiver, a sonar, a signal processing unit and a control unit. According to the data transmission and reception, it is mainly divided into two stages: data transmission stage and data reception stage.

In the data transmission stage, the transmitter obtains the information to be transmitted from the storage system of the underwater robot after being processed by the image processing module, which is referred to as secret information. Some information that can be publicly transmitted is used as the carrier of these processed information, which is referred to as public information.

The signal processing unit performs convolution coding on the public information to form carrier information; the secret information is also convolutionally coded, and then the secret information is interleaved through the interleaver. Finally, the secret information is embedded into the public information using a matrix embedding method based on binary Hamming code, which is called carrier information.

The secret information is encoded into a digital code by the signal source, a process also known as "data packing". In this process, a preamble and a guard interval are added to the data to enhance the reliability and robustness of the signal.

The coded data is modulated by the signal source onto a carrier frequency with orthogonal characteristics, a process also known as "signal modulation". The purpose of modulation is to allow data to be transmitted farther and more stably in the channel.

The modulated signal is sent out via sonar.

In the data receiving stage, the data signal from the transmitter is received by the receiver. The receiver demodulates the received signal, demodulates the signal from the carrier frequency, and restores it to the original data code. The demodulated data code is decoded by the receiver, and the preamble and synchronization code are removed to restore it to the original digital code or analog signal. What is obtained here is the secret information of the transmitter.

The secret information is first extracted by matrix quantization, so that the extracted information can be soft-decision Viterbi decoded to reduce the bit error rate of the secret information.

The result of matrix quantization extraction is deinterleaved and used as the input of Viterbi soft decision decoder, and finally decoded to recover the secret information.

Embodiment 2

As shown in FIG1 , the present invention also discloses an underwater target detection system based on covert communication technology and deep learning, including: an image processing module, an image acquisition module and a covert communication module;

The image acquisition module is used to acquire images of target objects in real time;

The image processing module is used to preprocess the image, and use a convolutional neural network model to extract features from the preprocessed image to obtain the bounding box coordinates, category label and confidence score of each target;

The covert communication module is used to achieve the secure transmission of the bounding box coordinates, category label and confidence score of each target and the communication information based on the covert communication information transmission method.

In this embodiment, as shown in FIG2 , for the image processing module (including: a preprocessing unit and a feature extraction unit), image data containing the target object is collected according to the requirements of the target detection task. The data can be obtained from a public data set, or it can be collected or purchased by itself. Ensure that the data set contains the diversity and representativeness of each category of targets.

Use LabelImg or RectLabel to annotate the collected images and mark the location and category of each target. Common target annotation formats include Bounding Box and Mask.

Divide the labeled data set into training set, validation set and test set. Usually use 70% to 80% of the data as training set, 10% to 15% of the data as validation set, and the remaining data as test set. Ensure that the target categories of each data set are relatively evenly distributed.

Apply data augmentation techniques to the training set to increase the diversity and quantity of the data set and improve the generalization ability of the model. Common data augmentation methods include random cropping, scaling, translation, rotation, brightness adjustment, color transformation, etc.

Convert annotation and image data into the format required for model training. For the YOLOv8 model, the commonly used data format is Darknet format or COCO format. You can use the corresponding tools to convert the dataset to the specified format.

During the training process, the image data is preprocessed, such as normalization, scaling to a fixed size, etc. This can speed up the training process and improve the performance and stability of the model.

Use the data loading tool DataLoader provided by the deep learning framework PyTorch to load the prepared dataset into the model for training.

The above is the data collection and preprocessing process, followed by the training process of a specific target detection model.

The network structure of the YOLOv8 model is built using the deep learning framework PyTorch. The construction process includes the combination of downsampling network, feature extraction layer and detection head.

Initialize the model weights randomly or with pre-trained weights.

Define the loss function of the YOLOv8 model. The loss function of YOLOv8 usually consists of position loss, classification loss, and confidence loss. You can use existing loss functions, such as cross entropy, and adjust them accordingly according to task requirements.

Use the data loading tool PyTorch's DataLoader to load the prepared training set and validation set data.

Select an appropriate optimizer, here we choose the SGD optimizer, and then set the corresponding hyperparameters such as learning rate, weight decay, momentum, etc. These hyperparameters can be debugged and optimized through experiments.

In each training cycle, a batch of training data is input into the model for forward propagation calculation, and then backpropagation is performed to update the model parameters. The model weights are updated according to the value of the loss function.

The rate of model training can be controlled by learning rate decay, which usually gradually reduces the learning rate during training.

Regularly save the model weights during the training process for subsequent model evaluation and inference.

Deploy the saved model to the operating system of the underwater robot for use in underwater target detection.

Underwater target detection requires images collected by underwater robots in real time. First, the images collected by underwater robots are preprocessed, including image size adjustment, pixel value normalization, etc. These preprocessing operations help improve the performance and accuracy of the model.

The convolutional neural network (CNN) model is used to extract features from the preprocessed images. YOLOv8 uses Darknet53 as the feature extraction network, which consists of multiple convolutional layers, residual blocks, and pooling layers.

YOLOv8 introduces multiple fine-grained features based on Darknet53 to extract feature information of different scales. By fusing features at different levels, the contextual information of objects of different sizes can be better captured.

YOLOv8 uses multiple detection heads to predict object boxes of different scales. Each detection head is responsible for predicting objects of a specific scale and outputs the corresponding classification probability, bounding box location, and confidence score.

The predictions output by the detection head are decoded to obtain the bounding box coordinates, category label, and confidence score of each object.

Since the same object may be detected by multiple bounding boxes, in order to eliminate duplicate detection results, the NMS algorithm is used to filter the predicted boxes. NMS selects the most likely target box based on the confidence score and overlap of the predicted boxes.

The target frame after NMS processing is further processed, such as setting a confidence threshold to filter out low-confidence target frames.

Finally, the output includes the category label, location information, and confidence score of the detected target box.

In this embodiment, as shown in FIG. 3 , the device for the image acquisition module mainly includes two major components: a high-definition camera module and an LED fill light device.

The underwater robot is remotely controlled to collect images of the inspection area and temporarily store the collected images in the storage system of the underwater robot.

In this embodiment, as shown in Figure 4, the communication module mainly includes the following components: a transmitter, a receiver, a sonar, a signal processing unit and a control unit. The data transmission and reception are mainly divided into two stages: a data transmission stage and a data reception stage.

In the data transmission stage, the transmitter obtains the information to be transmitted from the storage system of the underwater robot after being processed by the image processing module, which is referred to as secret information. Some information that can be publicly transmitted is used as the carrier of these processed information, which is referred to as public information.

The signal processing unit performs convolution coding on the public information to form carrier information; the secret information is also convolutionally coded, and then the secret information is interleaved through the interleaver. Finally, the secret information is embedded into the public information using a matrix embedding method based on binary Hamming code, which is called carrier information.

The secret information is encoded into a digital code by the signal source, a process also known as "data packing". In this process, a preamble and a guard interval are added to the data to enhance the reliability and robustness of the signal.

The coded data is modulated by the signal source onto a carrier frequency with orthogonal characteristics, a process also known as "signal modulation". The purpose of modulation is to allow data to be transmitted farther and more stably in the channel.

The modulated signal is sent out via sonar.

In the data receiving stage, the data signal from the transmitter is received by the receiver. The receiver demodulates the received signal, demodulates the signal from the carrier frequency, and restores it to the original data code. The demodulated data code is decoded by the receiver, and the preamble and synchronization code are removed to restore it to the original digital code or analog signal. What is obtained here is the secret information of the transmitter.

The secret information is first extracted by matrix quantization, so that the extracted information can be soft-decision Viterbi decoded to reduce the bit error rate of the secret information.

The result of matrix quantization extraction is deinterleaved and used as the input of Viterbi soft decision decoder, and finally decoded to recover the secret information.

The present invention mainly combines the relevant technologies in the field of target detection and covert communication, and integrates the two into the underwater robot. In the image processing module, a target detection algorithm based on yolov8 is adopted; in the communication module, an improved coding layer covert communication scheme is adopted.

Compared with the prior art, the technical solution of the present invention has the following advantages:

It can detect targets at a higher frame rate and has high real-time performance.

It adopts a deeper network structure and introduces multi-scale prediction and feature fusion strategies, which has better detection accuracy.

It is insensitive to changes in the number of targets and can better detect multi-target images.

Using interleaving in the covert communication system at the coding layer can effectively disperse the continuous errors in information extraction.

The addition of matrix quantization extraction can further reduce the bit error rate of secret information

Embodiment 3

The present invention also discloses an underwater target detection robot based on covert communication technology and deep learning, and the robot applies the underwater target detection method based on covert communication technology and deep learning. A schematic diagram of the mechanical shell of the underwater target detection robot is shown in FIG5 .

The embodiments described above are only descriptions of the preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Without departing from the design spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by ordinary technicians in this field should fall within the protection scope determined by the claims of the present invention.

Claims (8)

1. The underwater target detection method based on the hidden communication technology and the deep learning is characterized by comprising the following steps of:

acquiring an image of a target object in real time;

Preprocessing the image, and extracting features of the preprocessed image by using a convolutional neural network model to obtain boundary frame coordinates, category labels and confidence scores of each target;

The information transmission method based on the hidden communication realizes the safe transmission of the boundary frame coordinates, the category labels, the confidence scores and the communication information of each target;

The information transmission method based on the hidden communication, the method for realizing the safe transfer of the boundary frame coordinates, the category labels and the confidence scores of each target and the communication information comprises the following steps:

Performing convolutional encoding operation on the public information to form carrier information; carrying out convolutional encoding operation on the secret information, and carrying out interleaving treatment on the secret information through an interleaver; embedding secret information into public information by using a matrix embedding mode based on binary Hamming codes to obtain secret carrying information; modulating the secret information to obtain modulated signals and transmitting the modulated signals through sonar; wherein, the secret information is information processed by the image processing module;

Carrying out matrix quantization extraction on the secret information, de-interleaving the result of matrix quantization extraction as input of a Viterbi soft decision decoder, and decoding to recover the secret information;

The information transmission method based on the hidden communication realizes the safe transmission of the boundary frame coordinates, the category labels and the confidence scores of each target and the communication information, and mainly comprises the following components: a transmitter, a receiver, a sonar, a signal processing unit and a control unit; the data transmission and reception is mainly divided into two stages: a data transmission stage and a data reception stage;

In the data transmission stage, the transmitter acquires information which is to be transmitted and is processed by the image processing module from a storage system of the underwater robot, wherein the information is called secret information; the public transmission information is used as a carrier of the processed information and is called public information;

The signal processing unit carries out convolution encoding operation on the public information to form carrier information; the secret information is subjected to convolutional coding operation, and then is subjected to interleaving treatment through an interleaver; finally, embedding the secret information into the public information by using a matrix embedding mode based on binary Hamming codes, which is called as secret carrying information;

the secret information is encoded into digital codes by the signal source, and is called as data packing; the data is added with a preamble and a guard interval;

The coded data is modulated by a signal source onto a carrier frequency having orthogonal characteristics, referred to as "signal modulation";

the modulated signal is sent out through sonar;

In the data receiving stage, receiving a data signal from a transmitting end through a receiver; the receiver demodulates the received signal, demodulates the signal from the carrier frequency, and restores the signal into the original data code; the demodulated data code is decoded by a receiver, the lead code and the synchronous code are removed, and the original digital code or analog signal is recovered to obtain secret carrying information of a transmitting end;

Firstly, carrying out matrix quantization extraction on secret information, so that soft decision Viterbi decoding is carried out on the extracted information;

And the result of matrix quantization extraction is used as the input of a Viterbi soft decision decoder after de-interleaving, and finally the secret information is decoded and recovered.

2. The underwater target detection method based on the covert communication technology and the deep learning according to claim 1, wherein the method for acquiring the image of the target object in real time comprises the following steps:

And acquiring an image of the region to be detected by using the high-definition camera module to obtain an image of the target object.

3. The underwater target detection method based on the hidden communication technology and the deep learning according to claim 1, wherein the method for preprocessing the image and extracting features of the preprocessed image by using a convolutional neural network model to obtain the boundary frame coordinates, the category labels and the confidence scores of each target comprises the following steps:

Adjusting the size of the image of the target object and normalizing the pixel value;

based on YOLOv model, introducing a plurality of fine-grained feature layers on the basis of Darknet, extracting and fusing the feature information of different scales of the preprocessed image, and predicting target frames of different scales by using a plurality of detection heads to obtain the boundary frame coordinates, class labels and confidence scores of each target.

4. Underwater target detection system based on hidden communication technology and deep learning, which is characterized by comprising: the device comprises an image processing module, an image acquisition module and a hidden communication module;

the image acquisition module is used for acquiring an image of a target object in real time;

The image processing module is used for preprocessing the image, and extracting features of the preprocessed image by using a convolutional neural network model to obtain the boundary frame coordinates, the category labels and the confidence scores of each target;

The hidden communication module is used for realizing the safe transfer of the boundary frame coordinates, the category labels, the confidence scores and the communication information of each target based on the information transmission method of the hidden communication.

5. The underwater target detection system based on the covert communication technique and the deep learning of claim 4, wherein the image acquisition module comprises: the high-definition camera module and the LED light supplementing device;

The high-definition camera module is used for collecting images of the region to be detected and obtaining images of the target object;

The LED light supplementing device is used for assisting the high-definition camera module.

6. The underwater target detection system based on the covert communication technique and the deep learning of claim 4, wherein the image processing module comprises: a preprocessing unit and a feature extraction unit;

The preprocessing unit is used for adjusting the size of the image of the target object and normalizing the pixel value;

The feature extraction unit is used for introducing a plurality of fine-grained feature layers on the basis of a YOLOv model and Darknet, extracting and fusing feature information of different scales of the preprocessed image, and predicting target frames of different scales by using a plurality of detection heads to obtain the boundary frame coordinates, category labels and confidence scores of each target.

7. The underwater target detection system based on the covert communication technique and deep learning of claim 4, wherein the covert communication module comprises: a transmitter, a receiver, a sonar, a signal processing unit and a control unit;

The signal processing unit is used for performing convolution encoding operation on the public information to form carrier information; carrying out convolutional encoding operation on the secret information, and carrying out interleaving treatment on the secret information through an interleaver; embedding secret information into public information by using a matrix embedding mode based on binary Hamming codes to obtain secret carrying information; modulating the secret information to obtain a modulated signal; wherein, the secret information is information processed by the image processing module;

the transmitter is used for transmitting the secret carrying information;

The sonar is used for sending out the modulated signals;

The receiver is used for receiving the data signal from the transmitter;

The control unit is used for carrying out matrix quantization extraction on the secret information, de-interleaving the result of matrix quantization extraction, and then taking the result as the input of the Viterbi soft decision decoder, and decoding and recovering the secret information.

8. An underwater target detection robot based on a covert communication technology and deep learning, which is characterized in that the robot applies the underwater target detection method based on the covert communication technology and the deep learning as set forth in any one of claims 1 to 3.