CN117437595A - Fishing boat boundary crossing early warning method based on deep learning - Google Patents
Fishing boat boundary crossing early warning method based on deep learning Download PDFInfo
- Publication number
- CN117437595A CN117437595A CN202311598790.3A CN202311598790A CN117437595A CN 117437595 A CN117437595 A CN 117437595A CN 202311598790 A CN202311598790 A CN 202311598790A CN 117437595 A CN117437595 A CN 117437595A
- Authority
- CN
- China
- Prior art keywords
- fishing boat
- data
- fishing
- crossing
- detection
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 43
- 238000013135 deep learning Methods 0.000 title claims abstract description 22
- 238000001514 detection method Methods 0.000 claims abstract description 69
- 238000012544 monitoring process Methods 0.000 claims abstract description 20
- 238000005457 optimization Methods 0.000 claims abstract description 15
- 238000012549 training Methods 0.000 claims abstract description 14
- 238000002372 labelling Methods 0.000 claims abstract description 13
- 238000012795 verification Methods 0.000 claims abstract description 13
- 238000012360 testing method Methods 0.000 claims abstract description 9
- 238000012216 screening Methods 0.000 claims abstract description 5
- 238000004590 computer program Methods 0.000 claims description 13
- 230000006870 function Effects 0.000 claims description 11
- 238000003860 storage Methods 0.000 claims description 9
- 238000010200 validation analysis Methods 0.000 claims description 2
- 238000012545 processing Methods 0.000 description 5
- 239000011159 matrix material Substances 0.000 description 4
- 230000008901 benefit Effects 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 3
- 238000010606 normalization Methods 0.000 description 3
- 230000004913 activation Effects 0.000 description 2
- 238000004422 calculation algorithm Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000011218 segmentation Effects 0.000 description 2
- 230000003321 amplification Effects 0.000 description 1
- 238000003491 array Methods 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000003199 nucleic acid amplification method Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000011176 pooling Methods 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 230000001629 suppression Effects 0.000 description 1
- 238000013519 translation Methods 0.000 description 1
- 230000014616 translation Effects 0.000 description 1
- 238000011179 visual inspection Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/0464—Convolutional networks [CNN, ConvNet]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Evolutionary Computation (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Computing Systems (AREA)
- Software Systems (AREA)
- Multimedia (AREA)
- Life Sciences & Earth Sciences (AREA)
- Medical Informatics (AREA)
- Databases & Information Systems (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Molecular Biology (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Image Analysis (AREA)
Abstract
The invention provides a fishing boat boundary crossing early warning method based on deep learning, and belongs to the technical field of target detection. The method comprises the following steps: s1, acquiring fishing boat data, screening fishing boat images, and constructing fishing boat target detection data; s2, labeling the fishing boat data, generating image data with labels, and dividing the image data with the labels into a data set, a verification set and a test set; s3, constructing a fishing boat target detection optimization model, and training the model by utilizing a data set; s4, performing target detection on the fishing boat in the monitoring area based on the fishing boat target detection optimization model, outputting a classification result and a labeling frame, and performing multi-target tracking on the output classification result; s5, demarcating out-of-limit red lines in the monitoring area, and carrying out early warning prompt on fishing boats crossing out-of-limit red lines. The invention realizes the monitoring and boundary crossing early warning of the fishing boat in the video monitoring within 24 hours, can save a great deal of manual checking cost and effectively improves the monitoring efficiency of the fishing boat in the video monitoring.
Description
Technical Field
The invention relates to a fishing boat boundary crossing early warning method, in particular to a fishing boat boundary crossing early warning method based on deep learning, and belongs to the technical field of target detection.
Background
With the development of fishing industry, fishing boat data are increased increasingly, and how to carry out effective out statistics on the monitoring quantity of crossing the boundary on fishing boats in key areas of river in the boundary river becomes an urgent need to solve the problem. Especially fishing boat fishing occurs at night and the like, and uninterrupted continuous statistics are required. The existing fishing boat detection and boundary crossing early warning mainly adopts law enforcement ship on-site check supervision or camera deployment and radar manual visual inspection. The former law enforcement is inefficiency, and manual work and economic cost are high, and the latter needs artifical 24 hours to observe, the condition of leaking to examine the false detection easily appears.
Disclosure of Invention
The following presents a simplified summary of the invention in order to provide a basic understanding of some aspects of the invention. It should be understood that this summary is not an exhaustive overview of the invention. It is not intended to identify key or critical elements of the invention or to delineate the scope of the invention. Its purpose is to present some concepts in a simplified form as a prelude to the more detailed description that is discussed later.
In view of the above, the invention provides a fishing boat boundary crossing early warning method based on deep learning in order to solve the technical problems of high detection cost and poor effect in the prior art.
A fishing boat boundary crossing early warning method based on deep learning comprises the following steps:
s1, acquiring fishing boat data, screening fishing boat images, and constructing fishing boat target detection data;
s2, labeling the fishing boat data, generating image data with labels, and dividing the image data with the labels into a data set, a verification set and a test set;
s3, constructing a fishing boat target detection optimization model, and training the model by utilizing a data set;
the fishing boat target detection optimization model comprises an input end, a backbone network, a neck network and a detection layer;
the backbone network consists of four structures, namely a Focus module, a CBS module, a C2fGC module and an SSP module;
the C2fGC module comprises a GCBlock module, a Resunit module and a Concat module;
s4, performing target detection on the fishing boat in the monitoring area based on the fishing boat target detection optimization model, outputting a classification result and a labeling frame, and performing multi-target tracking on the output classification result;
s5, demarcating out-of-limit red lines in the monitoring area, and carrying out early warning prompt on fishing boats crossing out-of-limit red lines.
Preferably, labeling the fishing boat data, generating image data with a label, and dividing the image data with the label into a data set, a verification set and a test set comprises the following steps:
using LabelImg image marking tool to mark the rotated, segmented, amplified or translated image, wherein the two marked image and label are respectively images catalog and labels catalog;
the ratio of the division into data sets, validation sets and test sets may be according to 6:2:2 division.
Preferably, the image input format of the input terminal is 640×640×3.
Preferably, the training of the model using the dataset is as follows: defining training parameters and training round number, calculating a loss function, namely counter propagation gradient, and performing accuracy verification by using a verification set.
Preferably, the target detection is performed on the fishing boat in the monitoring area based on the fishing boat target detection optimization model, the classification result and the labeling frame are output, and the multi-target tracking method is performed on the output classification result, wherein the multi-target tracking method comprises the following steps:
s41, acquiring a detection frame and a corresponding detection score through a detector, classifying the detection frame into a high confidence coefficient group if the score is higher than T_high, and classifying the detection frame into a low confidence coefficient group if the score is lower than T_high and higher than T_low;
s42, reserving a high-confidence detection frame which is not matched with the track for the first time and a track which is not matched with the detection frame;
s43, associating the high-confidence detection frames which are not matched with the tracks and the tracks which are not matched with the detection frames, reserving the tracks which are not matched with the boundary frames for the second time, deleting the boundary frames which are not matched with the corresponding tracks after the second time of matching in the boundary frames with low confidence, and recognizing the boundary frames as the background which does not contain any object;
s44, storing the high-confidence boundary boxes which are not matched with the corresponding tracks as new tracks.
Preferably, the method further comprises S6, counting the fishing boats crossing the out-of-range red line, wherein the method comprises the following steps:
s61, setting a LINE crossing detection position in a video by using a Supervision method, setting a LINE starting point by LINE_START and a LINE ending point by LINE_END, and visually displaying the LINE crossing;
s62, acquiring a fishing boat ID identified by ByteTrack, and setting a fishing boat with a display confidence level of 0.3;
s63, prompting the fishing boat ID crossing the set line and counting the number;
s64, the stored out-of-range data is used for carrying out video interception on the out-of-range fishing boat and storing the video interception.
The second scheme is an electronic device, which comprises a memory and a processor, wherein the memory stores a computer program, and the processor realizes the first scheme of the fishing boat boundary crossing early warning method based on deep learning when executing the computer program.
According to a third aspect, a computer readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the method for warning of a fish-boat crossing boundary based on deep learning is implemented.
The beneficial effects of the invention are as follows: according to the invention, the visible light and infrared fishing boat data are acquired, the fishing boat is subjected to label manufacturing, the existing model is improved, the improved model is trained by using the fishing boat data marked, and the training result model is used for fishing boat target detection. And marking and tracking the ID of the fishing boat by using a ByteTrack multi-target tracker, demarcating a boundary line by using a Supervision method, and carrying out early warning and quantity statistics on the out-of-range fishing boat. The method can be used for carrying out high-precision identification and detection on local special fishing boats, tracking the fishing boats in the video monitored by the camera, and carrying out automatic early warning on the behavior of the fishing boats crossing the boundary. The method for deep learning can monitor the fishing boat in video monitoring and early warning the crossing of the boundary for 24 hours, can save a great deal of manual checking cost, and can effectively improve the monitoring efficiency of the fishing boat in video monitoring.
Drawings
The accompanying drawings, which are included to provide a further understanding of the invention and are incorporated in and constitute a part of this specification, illustrate embodiments of the invention and together with the description serve to explain the invention and do not constitute a limitation on the invention. In the drawings:
FIG. 1 is a flow chart of a method for early warning of a fish-boat crossing based on deep learning;
FIG. 2 is a schematic diagram of a fishing boat target detection optimization model structure;
FIG. 3 is a schematic view of a C2fGC module;
fig. 4 is a schematic view of a GCBlock module structure.
Detailed Description
In order to make the technical solutions and advantages of the embodiments of the present invention more apparent, the following detailed description of exemplary embodiments of the present invention is provided in conjunction with the accompanying drawings, and it is apparent that the described embodiments are only some embodiments of the present invention and not exhaustive of all embodiments. It should be noted that, without conflict, the embodiments of the present invention and features of the embodiments may be combined with each other.
Example 1: referring to fig. 1 to 4, the embodiment is described as a fishing boat boundary crossing early warning method based on deep learning, which includes the following steps:
s1, acquiring fishing boat data, screening fishing boat images, and constructing fishing boat target detection data;
acquiring fishing boat data: acquiring fishing boat image data from the COCO data set, the VOC data set and the image; screening fishing boat data from the COCO data set and the VOC data set, and intercepting the image data with the fishing boat from the video monitoring; the fishing boat data comprises visible light image data and infrared image data;
performing rotation, segmentation, amplification or translation processing on the fishing boat image so as to expand the sample size of the data set;
s2, labeling the fishing boat data, generating image data with labels, and dividing the image data with the labels into a data set, a verification set and a test set;
using LabelImg image marking tool to mark the rotated, segmented, amplified or translated image, wherein the two marked image and label are respectively images catalog and labels catalog; dividing the image into a data set, a verification set and a test set; the division ratio may be according to 6:2:2;
the invention improves the small target lifting of the fishing boat detection based on the original model of the deep learning target detection, and specifically comprises the following steps:
s3, constructing a fishing boat target detection optimization model, and training the model by utilizing a data set;
the fishing boat target detection optimization model comprises an input end, a backbone network, a neck network and a detection layer;
the input end carries out Mosaic data enhancement, self-adaptive anchor frame calculation and self-adaptive picture scaling operation; the image input format of the input end is 640×640×3;
the backbone network consists of four structures, namely a Focus module, a CBS module, a C2fGC module and an SSP module;
the Focus module performs slicing operation on the image; the CBS module comprises a Conv convolution layer, a batch normalization module and a SiLu activation function; the Conv convolution layer carries out convolution operation, and the batch normalization module carries out batch normalization processing;
the C2fGC module comprises a GCBlock module, a Resunit module and a Concat module; the structure of the C2fGC module is shown in FIG. 3, and the C2fGC module can improve the perceived information transmission efficiency of the model; the method replaces a convolution block in a traditional model with a GCBlock module, segments the convolution result, stacks the GCBlock after segmentation, and splices the convolution results of each layer together and convolves the result again;
GCBlock Module architecture referring to FIG. 4, the GCBlock module combines the ideas of a SimplifiedNLblock module and a SE block module. The method comprises the steps of inputting a feature map into a context modeling module (global attention pooling module), obtaining attention weights through a softmax function after the input passes through convolution W_k of 1*1, multiplying the weights with the original input, and enabling the weights to influence the input so as to obtain global features; in order to obtain the advantage of SE block light weight, the obtained global features are input into two layers of Bottleneck convolution, the BN layer is increased before the ReLU activation function to reduce the optimization difficulty, and regularization is introduced into the BN layer to increase the generalization capability of the model;
the neck network consists of an FPN feature pyramid and a PAN bottom-up feature enhancement algorithm;
the detection layer carries out 1 multiplied by 1 convolution operation on a 3-layer characteristic network generated by the neck network, and then outputs 3 groups of characteristic diagrams; the feature result carries out LOSS function calculation and non-maximum suppression through CIOU_LOSS+DIOU_nms, and a classification result and a feature frame are output;
CIOU_LOSS LOSS function:
wherein IOU is cross ratio, distance_2 2 distance_C representing the Euclidean Distance of the center points of the predicted and real frames 2 Represented is the diagonal distance of the minimum closure area that can contain both the predicted and real frames,v is a parameter measuring aspect ratio uniformity, defined as:
wherein w is p 、h p And w gt 、h gt Representing the height and width of the predicted frame and the height and width of the real frame respectively;
the method for training the model by utilizing the data set comprises the following steps: defining training parameters and training wheel numbers, performing back propagation gradient calculation through a loss function, performing accuracy verification by using a verification set, and finally converging the model to obtain a small target fishing boat detection model and a weight file with a format of pt.
S4, performing target detection on the fishing boat in the monitoring area based on the fishing boat target detection optimization model, outputting a classification result and a labeling frame, and performing multi-target tracking on the output classification result; the invention uses a ByteTrack multi-target tracker to label the classification result and the label frame, and simultaneously removes the predicted frame and the ID with the threshold value lower than 0.3 to track and monitor the targets of the multi-fishing boat. The multi-target fishing boat tracking method by ByteTrack is as follows:
s41, acquiring a detection frame and a corresponding detection score through a detector, classifying the detection frame into a high confidence coefficient group if the score is higher than T_high, and classifying the detection frame into a low confidence coefficient group if the score is lower than T_high and higher than T_low;
the matching process uses the similarity between the detection frame and the Kalman filtering estimation result, adopts IoU or Re-ID feature distance as similarity measurement, and then adopts Hungary algorithm to match based on the similarity, specifically: s42, reserving a high-confidence detection frame which is not matched with the track for the first time and a track which is not matched with the detection frame;
s43, associating the high-confidence detection frames which are not matched with the tracks and the tracks which are not matched with the detection frames, reserving the tracks which are not matched with the boundary frames for the second time, deleting the boundary frames which are not matched with the corresponding tracks after the second time of matching in the boundary frames with low confidence, and recognizing the boundary frames as the background which does not contain any object;
s44, storing the high-confidence boundary boxes which are not matched with the corresponding tracks as new tracks.
By using a Supervision method, setting out-of-limit red lines in continuous video monitoring, and carrying out early warning prompt and quantity statistics on fishing boats crossing the set red lines, wherein the method specifically comprises the following steps: s5, demarcating out-of-limit red lines in the monitoring area, and carrying out early warning prompt on fishing boats crossing out-of-limit red lines;
s6, counting fishing boats crossing the out-of-limit red line;
s61, setting a LINE crossing detection position in a video by using a Supervision method, setting a LINE starting point by LINE_START and a LINE ending point by LINE_END, and visually displaying the LINE crossing;
s62, acquiring a fishing boat ID identified by ByteTrack, and setting a fishing boat with a display confidence level of 0.3;
s63, prompting the fishing boat ID crossing the set line and counting the number;
s64, the stored out-of-range data is used for carrying out video interception on the out-of-range fishing boat and storing the video interception.
In order to evaluate the accuracy of the present invention, the present invention further comprises: s7, evaluating a target detection result of the fishing boat, wherein the method comprises the following steps:
judging whether the position of the predicted frame is accurate by adopting the cross ratio:
wherein B is p For prediction frame, B gt For the labeling frame, according to the relation with the labeling frame, dividing a certain prediction frame into one of the following four classes:
TP:{Conf＞P thresh and IOU > IOU thresh }
FP:{Conf＞P thresh And IOU < IOU thresh }
FN:{Conf＜P thresh And IOU > IOU thresh }
TN:{Conf＜P thresh And IOU < >IOU thresh }
In IOU thresh Constant between 0 and 1, IOU thresh Designated by the person; for a particular class, the number of TP, FP, FN, TN four prediction frames constitutes a Confusion Matrix (fusion Matrix); referring to table 1 confusion matrix table:
TABLE 1 confusion matrix table
| Number of predicted frames | Predicted as fishing boat | Prediction into other classes |
| Actually being a fishing boat | Number of TP prediction frames | Number of FN prediction frames |
| True to other categories | Number of FP prediction frames | TN prediction frame number |
Precision refers to the ratio of prediction accuracy among all prediction frames:
recall refers to the correctly predicted proportion of all label boxes:
average Precision (evaluation accuracy, AP for short): will P thresh The threshold value varies from 0 to 1, and each P is calculated thresh And drawing a PR performance curve of a certain class by Precision and Recall corresponding to the threshold value, wherein the enclosed area is the AP of the class. Take [email protected] as IOU thresh At 0.5, the value of AP is taken. [email protected] IOU is 0.95 thresh The average value of AP was taken as ten numbers increasing from 0.5 to 0.95 in steps of 0.05.
[email protected]＝AP i (IOU thresh ＝0.5)
[email protected]:0.95＝∑ j AP i (IOU thresh ＝j)
Based on an improved YOLOv8 detection model, IOU, precision, recall, AP, [email protected] and [email protected]:0.95 are used as performance indexes, and the accuracy of small target detection of the fishing boat can be improved by more than 5%.
Example 2: the computer device of the present invention may be a device including a processor and a memory, such as a single chip microcomputer including a central processing unit. And the processor is used for realizing the steps of the fishing boat boundary crossing early warning method based on deep learning when executing the computer program stored in the memory.
The processor may be a central processing unit (Central Processing Unit, CPU), other general purpose processors, digital signal processors (Digital Signal Processor, DSP), application specific integrated circuits (Application Specific Integrated Circuit, ASIC), off-the-shelf programmable gate arrays (Field-Programmable Gate Array, FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, or the like. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.
The memory may mainly include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application program (such as a sound playing function, an image playing function, etc.) required for at least one function, and the like; the storage data area may store data (such as audio data, phonebook, etc.) created according to the use of the handset, etc. In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as a hard disk, memory, plug-in hard disk, smart Media Card (SMC), secure Digital (SD) Card, flash Card (Flash Card), at least one disk storage device, flash memory device, or other volatile solid-state storage device.
Example 3: computer-readable storage medium embodiments.
The computer readable storage medium of the present invention may be any form of storage medium that is read by a processor of a computer device, including but not limited to a nonvolatile memory, a volatile memory, a ferroelectric memory, etc., on which a computer program is stored, and when the processor of the computer device reads and executes the computer program stored in the memory, the steps of a fishing boat boundary crossing early warning method based on deep learning described above may be implemented.
The computer program comprises computer program code which may be in source code form, object code form, executable file or in some intermediate form, etc. The computer readable medium may include: any entity or device capable of carrying the computer program code, a recording medium, a U disk, a removable hard disk, a magnetic disk, an optical disk, a computer Memory, a Read-Only Memory (ROM), a random access Memory (RAM, random Access Memory), an electrical carrier signal, a telecommunications signal, a software distribution medium, and so forth. It should be noted that the computer readable medium contains content that can be appropriately scaled according to the requirements of jurisdictions in which such content is subject to legislation and patent practice, such as in certain jurisdictions in which such content is subject to legislation and patent practice, the computer readable medium does not include electrical carrier signals and telecommunication signals.
While the invention has been described with respect to a limited number of embodiments, those skilled in the art, having benefit of the above description, will appreciate that other embodiments are contemplated within the scope of the invention as described herein. Furthermore, it should be noted that the language used in the specification has been principally selected for readability and instructional purposes, and may not have been selected to delineate or circumscribe the inventive subject matter. Accordingly, many modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the appended claims. The disclosure of the present invention is intended to be illustrative, but not limiting, of the scope of the invention, which is defined by the appended claims.
Claims (8)
1. A fishing boat boundary crossing early warning method based on deep learning is characterized by comprising the following steps:
s1, acquiring fishing boat data, screening fishing boat images, and constructing fishing boat target detection data;
s2, labeling the fishing boat data, generating image data with labels, and dividing the image data with the labels into a data set, a verification set and a test set;
s3, constructing a fishing boat target detection optimization model, and training the model by utilizing a data set;
the fishing boat target detection optimization model comprises an input end, a backbone network, a neck network and a detection layer;
the backbone network consists of four structures, namely a Focus module, a CBS module, a C2fGC module and an SSP module;
the C2fGC module comprises a GCBlock module, a Resunit module and a Concat module;
s4, performing target detection on the fishing boat in the monitoring area based on the fishing boat target detection optimization model, outputting a classification result and a labeling frame, and performing multi-target tracking on the output classification result;
s5, demarcating out-of-limit red lines in the monitoring area, and carrying out early warning prompt on fishing boats crossing out-of-limit red lines.
2. The method for pre-warning the boundary crossing of the fishing boat based on the deep learning according to claim 1, wherein the method for labeling the fishing boat data, generating the image data with the label and dividing the image data with the label into a data set, a verification set and a test set is as follows:
using LabelImg image marking tool to mark the rotated, segmented, amplified or translated image, wherein the two marked image and label are respectively images catalog and labels catalog;
the ratio of division into data set, validation set and test set is according to 6:2:2 division.
3. The method for pre-warning of a fish-boat crossing based on deep learning as claimed in claim 1, wherein the image input format of the input end is 640 x 3.
4. The method for pre-warning of a fish-boat crossing based on deep learning as claimed in claim 1, wherein the training of the model by using the data set is: defining training parameters and training round number, calculating a loss function, namely counter propagation gradient, and performing accuracy verification by using a verification set.
5. The method for pre-warning the boundary crossing of the fishing vessel based on the deep learning according to claim 1, wherein the method for multi-objective tracking of the outputted classification result by performing objective detection on the fishing vessel in the monitoring area based on the fishing vessel objective detection optimization model and outputting the classification result and the marking frame is as follows:
s41, acquiring a detection frame and a corresponding detection score through a detector, classifying the detection frame into a high confidence coefficient group if the score is higher than T_high, and classifying the detection frame into a low confidence coefficient group if the score is lower than T_high and higher than T_low;
s42, reserving a high-confidence detection frame which is not matched with the track for the first time and a track which is not matched with the detection frame;
s43, associating the high-confidence detection frames which are not matched with the tracks and the tracks which are not matched with the detection frames, reserving the tracks which are not matched with the boundary frames for the second time, deleting the boundary frames which are not matched with the corresponding tracks after the second time of matching in the boundary frames with low confidence, and recognizing the boundary frames as the background which does not contain any object;
s44, storing the high-confidence boundary boxes which are not matched with the corresponding tracks as new tracks.
6. The method for pre-warning the boundary crossing of the fishing boat based on the deep learning as set forth in claim 1, further comprising, S6, counting the fishing boat crossing the boundary crossing red line, wherein the method comprises the following steps:
s61, setting a LINE crossing detection position in a video by using a Supervision method, setting a LINE starting point by LINE_START and a LINE ending point by LINE_END, and visually displaying the LINE crossing;
s62, acquiring a fishing boat ID identified by ByteTrack, and setting a fishing boat with a display confidence level of 0.3;
s63, prompting the fishing boat ID crossing the set line and counting the number;
s64, the stored out-of-range data is used for carrying out video interception on the out-of-range fishing boat and storing the video interception.
7. An electronic device comprising a memory and a processor, the memory storing a computer program, the processor implementing the steps of a deep learning based fishing vessel boundary crossing warning method of any one of claims 1-6 when executing the computer program.
8. A computer readable storage medium, characterized in that it has stored thereon a computer program, which when executed by a processor, implements a deep learning based fishing vessel boundary crossing warning method according to any of claims 1-6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311598790.3A CN117437595A (en) | 2023-11-27 | 2023-11-27 | Fishing boat boundary crossing early warning method based on deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311598790.3A CN117437595A (en) | 2023-11-27 | 2023-11-27 | Fishing boat boundary crossing early warning method based on deep learning |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117437595A true CN117437595A (en) | 2024-01-23 |
Family
ID=89549927
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311598790.3A Pending CN117437595A (en) | 2023-11-27 | 2023-11-27 | Fishing boat boundary crossing early warning method based on deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117437595A (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113269073A (en) * | 2021-05-19 | 2021-08-17 | 青岛科技大学 | Ship multi-target tracking method based on YOLO V5 algorithm |
| CN115240119A (en) * | 2022-08-10 | 2022-10-25 | 哈尔滨航天恒星数据***科技有限公司 | Pedestrian small target detection method in video monitoring based on deep learning |
| CN115994953A (en) * | 2023-02-15 | 2023-04-21 | 国网浙江省电力有限公司电力科学研究院 | Power field security monitoring and tracking method and system |
| WO2023116631A1 (en) * | 2021-12-24 | 2023-06-29 | 中国科学院深圳先进技术研究院 | Training method and training apparatus for rotating-ship target detection model, and storage medium |
| CN116563478A (en) * | 2022-01-27 | 2023-08-08 | ***通信有限公司研究院 | Synchronous positioning and mapping SLAM algorithm, terminal and storage medium |
| CN116935332A (en) * | 2023-03-30 | 2023-10-24 | 华南理工大学 | Fishing boat target detection and tracking method based on dynamic video |
-
2023
- 2023-11-27 CN CN202311598790.3A patent/CN117437595A/en active Pending
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113269073A (en) * | 2021-05-19 | 2021-08-17 | 青岛科技大学 | Ship multi-target tracking method based on YOLO V5 algorithm |
| WO2023116631A1 (en) * | 2021-12-24 | 2023-06-29 | 中国科学院深圳先进技术研究院 | Training method and training apparatus for rotating-ship target detection model, and storage medium |
| CN116563478A (en) * | 2022-01-27 | 2023-08-08 | ***通信有限公司研究院 | Synchronous positioning and mapping SLAM algorithm, terminal and storage medium |
| CN115240119A (en) * | 2022-08-10 | 2022-10-25 | 哈尔滨航天恒星数据***科技有限公司 | Pedestrian small target detection method in video monitoring based on deep learning |
| CN115994953A (en) * | 2023-02-15 | 2023-04-21 | 国网浙江省电力有限公司电力科学研究院 | Power field security monitoring and tracking method and system |
| CN116935332A (en) * | 2023-03-30 | 2023-10-24 | 华南理工大学 | Fishing boat target detection and tracking method based on dynamic video |
Non-Patent Citations (1)
| Title |
|---|
| 王炳德等: "基于YOLOv3 的船舶目标检测算法", 中国航海, vol. 43, no. 1, 25 March 2020 (2020-03-25) * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110047095B (en) | Tracking method and device based on target detection and terminal equipment | |
| CN111210399B (en) | Imaging quality evaluation method, device and equipment | |
| CN111175318A (en) | Screen scratch fragmentation detection method and equipment | |
| CN112836668A (en) | Ship target detection method, terminal device and storage medium | |
| CN113191353A (en) | Vehicle speed determination method, device, equipment and medium | |
| CN116503399B (en) | Insulator pollution flashover detection method based on YOLO-AFPS | |
| CN112634368A (en) | Method and device for generating space and OR graph model of scene target and electronic equipment | |
| CN111539456B (en) | Target identification method and device | |
| CN114462469B (en) | Training method of target detection model, target detection method and related device | |
| CN112288701A (en) | Intelligent traffic image detection method | |
| CN116824516A (en) | Road construction safety monitoring and management system | |
| CN110298302B (en) | Human body target detection method and related equipment | |
| CN116863286A (en) | Double-flow target detection method and model building method thereof | |
| CN111553183A (en) | Ship detection model training method, ship detection method and ship detection device | |
| CN114708291A (en) | Image processing method, image processing device, electronic equipment and storage medium | |
| CN116805387B (en) | Model training method, quality inspection method and related equipment based on knowledge distillation | |
| CN113205510A (en) | Railway intrusion foreign matter detection method, device and terminal | |
| CN112287905A (en) | Vehicle damage identification method, device, equipment and storage medium | |
| CN115953746B (en) | Ship monitoring method and device | |
| CN112884866A (en) | Coloring method, device, equipment and storage medium for black and white video | |
| CN116977895A (en) | Stain detection method and device for universal camera lens and computer equipment | |
| CN115346169B (en) | Method and system for detecting sleep post behaviors | |
| CN117437595A (en) | Fishing boat boundary crossing early warning method based on deep learning | |
| CN107886049B (en) | Visibility recognition early warning method based on camera probe | |
| CN114550129B (en) | Machine learning model processing method and system based on data set |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |