CN116310597A - Garbage classification and positioning method, unmanned cleaning boat control method and system - Google Patents

Garbage classification and positioning method, unmanned cleaning boat control method and system Download PDF

Info

Publication number
CN116310597A
CN116310597A CN202310515917.4A CN202310515917A CN116310597A CN 116310597 A CN116310597 A CN 116310597A CN 202310515917 A CN202310515917 A CN 202310515917A CN 116310597 A CN116310597 A CN 116310597A
Authority
CN
China
Prior art keywords
boat
garbage
unmanned cleaning
controlling
floating
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
Application number
CN202310515917.4A
Other languages
Chinese (zh)
Inventor
鲁仁全
张朗
徐雍
刘畅
彭慧
饶红霞
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Guangdong University of Technology
Original Assignee
Guangdong University of Technology
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Guangdong University of Technology filed Critical Guangdong University of Technology
Priority to CN202310515917.4A priority Critical patent/CN116310597A/en
Publication of CN116310597A publication Critical patent/CN116310597A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • 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
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B63—SHIPS OR OTHER WATERBORNE VESSELS; RELATED EQUIPMENT
    • B63B—SHIPS OR OTHER WATERBORNE VESSELS; EQUIPMENT FOR SHIPPING 
    • B63B35/00—Vessels or similar floating structures specially adapted for specific purposes and not otherwise provided for
    • B63B35/32—Vessels or similar floating structures specially adapted for specific purposes and not otherwise provided for for collecting pollution from open water
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S19/00—Satellite radio beacon positioning systems; Determining position, velocity or attitude using signals transmitted by such systems
    • G01S19/38—Determining a navigation solution using signals transmitted by a satellite radio beacon positioning system
    • G01S19/39—Determining a navigation solution using signals transmitted by a satellite radio beacon positioning system the satellite radio beacon positioning system transmitting time-stamped messages, e.g. GPS [Global Positioning System], GLONASS [Global Orbiting Navigation Satellite System] or GALILEO
    • G01S19/42—Determining position
    • 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
    • G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
    • G06V10/774—Generating sets of training patterns; Bootstrap methods, e.g. bagging or boosting
    • 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
    • B—PERFORMING OPERATIONS; TRANSPORTING
    • B63—SHIPS OR OTHER WATERBORNE VESSELS; RELATED EQUIPMENT
    • B63B—SHIPS OR OTHER WATERBORNE VESSELS; EQUIPMENT FOR SHIPPING 
    • B63B35/00—Vessels or similar floating structures specially adapted for specific purposes and not otherwise provided for
    • B63B2035/006—Unmanned surface vessels, e.g. remotely controlled
    • B63B2035/007—Unmanned surface vessels, e.g. remotely controlled autonomously operating
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
    • Y02A20/00—Water conservation; Efficient water supply; Efficient water use
    • Y02A20/20—Controlling water pollution; Waste water treatment
    • Y02A20/204—Keeping clear the surface of open water from oil spills

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Evolutionary Computation (AREA)
  • General Physics & Mathematics (AREA)
  • Health & Medical Sciences (AREA)
  • Computing Systems (AREA)
  • General Health & Medical Sciences (AREA)
  • Software Systems (AREA)
  • Artificial Intelligence (AREA)
  • Medical Informatics (AREA)
  • Databases & Information Systems (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Multimedia (AREA)
  • Remote Sensing (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Data Mining & Analysis (AREA)
  • Molecular Biology (AREA)
  • General Engineering & Computer Science (AREA)
  • Mathematical Physics (AREA)
  • Environmental & Geological Engineering (AREA)
  • Public Health (AREA)
  • Chemical & Material Sciences (AREA)
  • Combustion & Propulsion (AREA)
  • Mechanical Engineering (AREA)
  • Ocean & Marine Engineering (AREA)
  • Biomedical Technology (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Image Analysis (AREA)

Abstract

The invention belongs to the field of image recognition, and discloses a garbage classification and positioning method, an unmanned cleaning boat control method and a system, wherein the garbage classification and positioning method comprises the steps of acquiring a water surface image containing floating garbage; and inputting the water surface image into a trained improved yolov5 model to identify, and obtaining the type and the position of the floating garbage, wherein the improved yolov5 model comprises a backbox part containing a Resnet network, and a CBAM module is inserted after each block of the Resnet network. According to the garbage classification and positioning method, the water surface image containing the floating garbage is identified by adopting the trained improved yolov5 model based on a one-step method, so that the classification and the position of the floating garbage can be obtained through one-time operation. Therefore, the time for classifying and positioning the floating garbage on the water surface is saved, and the recovery efficiency is improved.

Description

Garbage classification and positioning method, unmanned cleaning boat control method and system
Technical Field
The invention relates to garbage recovery, in particular to a garbage classification and positioning method, an unmanned cleaning boat control method and an unmanned cleaning boat control system.
Background
Traditional water surface garbage classification is generally based on manual classification, and the classification mode is low in efficiency of classifying the water surface garbage. In order to improve the classification recovery efficiency of the water surface garbage, the unmanned clean boat is introduced in the prior art to classify and recover the water surface garbage, but the existing unmanned clean boat generally has no real-time classification function in the process of recovering the water surface garbage and can only directly landfill, burn or land the garbage again, so that the garbage treatment time is longer, and the recovery efficiency is influenced.
Disclosure of Invention
The invention aims to provide a garbage classification and positioning method, an unmanned cleaning boat control method and an unmanned cleaning boat control system, which solve the problem of improving the efficiency of classifying water surface garbage when classifying and recycling the water surface garbage by adopting an unmanned cleaning boat, thereby improving the recycling efficiency.
In order to achieve the above purpose, the present invention provides the following technical solutions:
in a first aspect, the present invention provides a garbage classification and positioning method, comprising:
acquiring a water surface image containing floating garbage;
inputting the water surface image into a trained improved yolov5 model for recognition to obtain the type and position of the floating garbage,
wherein the improved yolov5 model comprises a backbond part, the backbond part comprises a Resnet network, a Batchnormalization layer and a dropout layer are added into the Resnet network, a CBAM module is inserted after each block of the Resnet network,
wherein, add the batch normalization layer and the dropout layer in the Resnet network, include:
a batch normalization layer is added after the convolutional layer of each residual block and identity block, and a dropout layer is added before the last full-join layer.
Optionally, before acquiring the water surface image containing the floating garbage, the method further comprises:
improving the yolov5 model to obtain an improved yolov5 model;
and training the improved yolov5 model to obtain the trained improved yolov5 model.
Optionally, the modified yolov5 model further includes a neg portion and a Head portion.
Optionally, training the improved yolov5 model includes:
setting classification types of the water surface garbage;
acquiring a data set for training;
performing enhancement processing on the data set to obtain an enhanced data set;
dividing the enhanced data set into a training set, a verification set and a test set according to a set proportion;
performing label coding on the training set, the verification set and the test set;
selecting a loss function, an activation function, an optimizer, an evaluation function and a batch size respectively;
and training the improved yolov5 model by using the training set, the verification set and the test set after label coding to obtain the trained improved yolov5 model.
In a second aspect, the present invention provides a method of controlling an unmanned cleaning boat, comprising:
obtaining the type and position of floating garbage according to the garbage classification and positioning method of any one of the above;
controlling the unmanned cleaning boat to go to the position, wherein the unmanned cleaning boat comprises a front boat and a rear boat, and the front boat and the rear boat are designed in a separable way;
controlling the unmanned cleaning boat to salvage the floating garbage;
and conveying the floating garbage to the position corresponding to the rear boat according to the type of the floating garbage.
Optionally, after the floating garbage is transferred to the position corresponding to the rear boat, the method further comprises:
detecting whether the stored garbage amount in the rear boat reaches a warning line, if so, sending GPS positioning to a standby boat which does not store floating garbage; the standby boat is used for going to the position corresponding to the GPS positioning after the GPS positioning is received;
after the spare boat reaches the position corresponding to the GPS positioning, controlling the separation of the front boat and the rear boat;
controlling the front boat and the standby boat to be connected to form a new unmanned cleaning boat;
and controlling the rear boat to return to the shore.
In a third aspect, the invention provides an unmanned cleaning boat control system, which comprises an upper computer and an unmanned cleaning boat, wherein the unmanned cleaning boat comprises a front boat and a rear boat, and the front boat and the rear boat are designed in a separable way;
the unmanned cleaning boat further comprises a camera, a control panel, a garbage collection module, a conveyor belt and a garbage classification platform;
the camera, the control panel and the upper computer are mutually matched, and the type and the position of the floating garbage are obtained according to the garbage classification and positioning method;
the control panel is also used for controlling the unmanned cleaning boat to go to the position;
the control panel is also used for controlling the garbage collection module to salvage the floating garbage;
the control panel is also used for controlling the conveyor belt to transmit the floating garbage to the garbage classification platform;
the control panel is also used for controlling the garbage classification platform to transmit the floating garbage to the position corresponding to the rear boat according to the type of the floating garbage.
Optionally, the unmanned cleaning boat further comprises a baffle for concentrating the waste on the waste sorting deck.
Optionally, the rear boat comprises a detection module, a garbage recycling bin and a sending module;
the garbage recycling bin comprises a plurality of bins, and different bins are used for storing different types of floating garbage;
the detection module is used for detecting whether the stored garbage amount in the rear boat reaches a warning line or not;
the sending module is used for sending GPS positioning to the spare boat which does not store floating garbage when the stored garbage amount in the rear boat reaches a warning line; the standby boat is used for going to the position corresponding to the GPS positioning after the GPS positioning is received;
after the spare boat reaches the position corresponding to the GPS positioning, the upper computer is used for controlling the separation of the front boat and the rear boat;
the upper computer is also used for controlling the connection of the front boat and the standby boat to form a new unmanned cleaning boat;
the upper computer is also used for controlling the back boat to return to the shore.
Optionally, the structure of the spare boat is identical to the structure of the rear boat.
According to the garbage classification and positioning method, the water surface image containing the floating garbage is identified by adopting the trained improved yolov5 model based on a one-step method, so that the classification and the position of the floating garbage can be obtained through one-time operation. Therefore, the time for classifying and positioning the floating garbage on the water surface is saved, and the recovery efficiency is improved.
The invention applies the neural network based on deep learning, improves the existing garbage classification and classification algorithm, improves the information extraction capability of the Backbone, and improves the classification confidence of the optimized yolov5 Backbone network to 96 percent, thereby improving the success rate of one-time operation recognition.
Meanwhile, the garbage classification and positioning algorithm is applied to the unmanned cleaning boat, so that the unmanned cleaning boat can classify garbage while collecting the garbage, and the cost is saved.
According to the unmanned cleaning boat control method and the unmanned cleaning boat control system, the separable unmanned cleaning boat structure is designed, a standby rear boat can be replaced after garbage is fully loaded, and the time required for returning to and fro for cleaning the warehouse is reduced. Compared with the traditional unmanned cleaning boat, the method is more flexible in work, can effectively improve the working efficiency and expands the working range.
Drawings
The accompanying drawings, which are included to provide a further understanding of embodiments 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 principles of the invention.
FIG. 1 is a schematic diagram of the garbage classification and positioning method of the present invention.
FIG. 2 is a schematic diagram of a CBAM module of the present invention.
FIG. 3 is a schematic representation of the training of the improved yolov5 model of the present invention.
FIG. 4 is a schematic representation of the confidence of the improved yolov5 model of the present invention.
FIG. 5 is a schematic representation of the Loss curves of the training set and verification set of the present invention.
Fig. 6 is a schematic illustration of the unmanned cleaning boat control process of the present invention.
Fig. 7 is a schematic view of the unmanned cleaning boat of the present invention.
Detailed Description
In order that the above-recited objects, features and advantages of the present invention will be more clearly understood, a more particular description of the invention will be rendered by reference to the appended drawings and appended detailed description. In addition, the embodiments of the present invention and the features in the embodiments may be combined with each other without collision. In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present invention, however, the present invention may be practiced in other ways than within the scope of the description, and the scope of the invention is therefore not limited to the specific embodiments disclosed below.
Example 1:
in one embodiment as shown in fig. 1, the present invention provides a garbage classification and positioning method, including:
acquiring a water surface image containing floating garbage;
inputting the water surface image into a trained improved yolov5 model for recognition to obtain the type and position of the floating garbage,
the improved yolov5 model comprises a backbox part, the backbox part comprises a Resnet network, a Batchnormalization layer and a dropout layer are added into the Resnet network, and a CBAM module is inserted after each block of the Resnet network.
Wherein, add the batch normalization layer and the dropout layer in the Resnet network, include:
a batch normalization layer is added after the convolutional layer of each residual block and identity block, and a dropout layer is added before the last full-join layer.
Specifically, the Batchnormalization layer is used to batch normalize the data. The normalization makes the data conform to the distribution with the mean value of 0 and 1 as the standard deviation, and the batch normalization is to perform the normalization treatment on a small batch of data (batch). Batch normalization can alleviate the overfitting problem and gradient explosion/vanishing problem. The batch normalization formula is:
Figure SMS_1
Figure SMS_2
for data before mapping +.>
Figure SMS_3
For mapped data, ++>
Figure SMS_4
For this batch of data mean, +.>
Figure SMS_5
For this batch of data variance, constant +.>
Figure SMS_6
Ensuring that the denominator is not 0.
Specifically, a dropout layer is added in the network, and a part of neurons are frozen randomly each time, so that the problem of overfitting caused by excessive parameters is reduced.
Specifically, a CBAM module is inserted after each block of the Resnet 50. The network's characterizations are increased by an attention mechanism: important features are of interest, unnecessary features are suppressed. Because CBAM is lightweight and versatile, it can be seamlessly integrated into any CNN network.
The corresponding model of the CBAM module is shown in fig. 2.
According to the garbage classification and positioning method, the water surface image containing the floating garbage is identified by adopting the trained improved yolov5 model based on a one-step method, so that the classification and the position of the floating garbage can be obtained through one-time operation. Therefore, the time for classifying and positioning the floating garbage on the water surface is saved, and the recovery efficiency is improved.
The invention applies the neural network based on deep learning, improves the existing garbage classification and classification algorithm, improves the information extraction capability of the Backbone, and improves the classification confidence of the optimized yolov5 Backbone network to 96 percent, thereby improving the success rate of one-time operation recognition.
Optionally, before acquiring the water surface image containing the floating garbage, the method further comprises:
improving the yolov5 model to obtain an improved yolov5 model;
and training the improved yolov5 model to obtain the trained improved yolov5 model.
Optionally, the modified yolov5 model further includes a neg portion and a Head portion.
Optionally, as shown in fig. 3, training the improved yolov5 model includes:
setting classification types of the garbage on the water surface.
In one embodiment, classification is also different from daily garbage classification, as the water garbage is different from household garbage at ordinary times. According to the garbage treatment method, the water garbage is divided into three types: garbage, perishable garbage, other garbage can be recovered.
Perishable species: the agricultural production is easy to rot garbage such as water plants, weeds, dead branches and rotten leaves, melons and fruits, vegetable leaves, animal carcasses and crops.
Recoverable classes: glass bottles, beverage cans, plastic bottles, plastic (nylon) bags, foam plastic and other products suitable for recycling and resource utilization.
Other waste (non-perishable): bamboo-like wood products, branches, plates, polluted waste clothes, other textiles, packaging bags and other non-putrescible waste.
A data set for training is acquired.
In one embodiment, the data set collection approach is: picking from existing datasets, picking on the web, shooting itself, etc.
And carrying out enhancement processing on the data set to obtain an enhanced data set.
Specifically, the training database is enlarged by data enhancement of means such as image rotation, clipping and mirroring.
The enhanced data set is divided into a training set, a verification set and a test set according to a set proportion.
In one embodiment, all enhanced data sets are combined at 0.6:0.2: the ratio of 0.2 is divided into training set, validation set and test set.
And performing label coding on the training set, the verification set and the test set.
In one embodiment, since the water garbage is classified as a multi-classification problem, but with fewer categories, the tags are encoded as one-hot matrices, classification vectors.
The loss function, activation function, optimizer, evaluation function, and batch size are selected separately.
In one embodiment, selecting the loss function includes: since the tag is encoded as a one-hot matrix, the loss function used at this time should be: categorical_cross sentropy.
The loss function mathematical expression is:
Figure SMS_7
the Loss is the Loss of the Loss,
Figure SMS_8
an i-th value of the output vector that is Softmax, representing a probability of the i-th class; outputsize is the number of categories；/>
Figure SMS_9
A tag value representing a sample of the actual class i.
In one embodiment, selecting an activation function includes:
since the activation function softmax is more suitable for solving the multi-classification problem, the water garbage classification uses the softmax function as the activation function. The mathematical formula is as follows:
Figure SMS_10
if n classification problem, input a vector of n 1, then x i As the ith number in the vector, denominator
Figure SMS_11
X represents e j The sum of powers j is from 1 to n, the result being the probability of class i.
Figure SMS_12
The probability of the i-th class is represented, and e represents a natural constant.
In one embodiment, a selection optimizer includes:
from experiments one by one in various optimizations, the effect of the optimizer SGD (random gradient descent) is relatively good, and oscillation after convergence is small, so that the SGD is selected as the optimizer. It estimates the current gradient one sample at a time, and updates the model parameters once. The SGD optimizer formula is as follows:
Figure SMS_13
Figure SMS_14
for pre-update weight ++>
Figure SMS_15
For updating the post-weight +.>
Figure SMS_16
For learning rate->
Figure SMS_17
For->
Figure SMS_18
Gradient of->
Figure SMS_19
Is an SGD optimizer.
In one embodiment, selecting an evaluation function includes: as the garbage on water is classified into a multi-classification single-label task, and the label codes are one-hot, the evaluation index selects the category_accuracies.
In one embodiment, selecting a batch size includes:
batch size, i.e., batch_size, if too large, requires more time, consuming more memory capacity; too small a Batch size may be difficult to converge, so a moderate Batch size is required. The current training batch_size=16.
And training the improved yolov5 model by using the training set, the verification set and the test set after label coding to obtain the trained improved yolov5 model.
In one embodiment, the confidence curve of the trained modified yolov5 model is shown in FIG. 4, and the training set and validation set Loss curves are shown in FIG. 5.
As can be seen from fig. 4, the confidence of the training model reaches about 95%, which is enough to replace manual classification.
Example 2:
the invention provides a control method of an unmanned cleaning boat, which comprises the following steps:
obtaining the type and position of floating garbage according to the garbage classification and positioning method of any one of the above;
controlling the unmanned cleaning boat to go to the position, wherein the unmanned cleaning boat comprises a front boat and a rear boat, and the front boat and the rear boat are designed in a separable way;
controlling the unmanned cleaning boat to salvage the floating garbage;
and conveying the floating garbage to the position corresponding to the rear boat according to the type of the floating garbage.
Optionally, after the floating garbage is transferred to the position corresponding to the rear boat, the method further comprises:
detecting whether the stored garbage amount in the rear boat reaches a warning line, if so, sending GPS positioning to a standby boat which does not store floating garbage; the standby boat is used for going to the position corresponding to the GPS positioning after the GPS positioning is received;
after the spare boat reaches the position corresponding to the GPS positioning, controlling the separation of the front boat and the rear boat;
controlling the front boat and the standby boat to be connected to form a new unmanned cleaning boat;
and controlling the rear boat to return to the shore.
Specifically, the front boat, the rear boat and the spare boat are respectively represented by the boat A, the boat B and the boat C, and the garbage amount of the rear boat reaches a warning line, namely, after the rear boat is fully loaded, the specific control process is shown in fig. 6.
Example 3:
the invention provides an unmanned cleaning boat control system, which comprises an upper computer and an unmanned cleaning boat, wherein the unmanned cleaning boat comprises a front boat 1 and a rear boat 2, and the front boat 1 and the rear boat 2 are designed in a separable way;
as shown in fig. 7, the unmanned cleaning boat further comprises a camera 3, a control panel, a garbage collection module 4, a conveyor belt 5 and a garbage classification platform 7;
the place ahead of preceding ship 1 is provided with garbage collection module 4, and the top of preceding ship 1 is provided with camera 3, and conveyer belt 5 is located between preceding ship 1 and the back ship 2, and garbage classification platform 7 sets up the inside at back ship 2.
The camera 3, the control panel and the upper computer are mutually matched, and the type and the position of the floating garbage are obtained according to the garbage classification and positioning method;
the control panel is also used for controlling the unmanned cleaning boat to go to the position;
the control panel is also used for controlling the garbage collection module 4 to salvage the floating garbage;
the control panel is also used for controlling the conveyor belt 5 to transmit floating garbage to the garbage classification platform 7;
the control panel is also used for controlling the garbage classification platform 7 to transmit the floating garbage to the position corresponding to the rear boat 2 according to the type of the floating garbage.
In one embodiment, the camera 3, the control panel and the upper computer are mutually matched, and the type and the position of the floating garbage are obtained according to the garbage classification and positioning method described in any one of the above, which comprises the following steps:
the upper computer is used for sending the trained improved yolov5 model to the control panel;
the camera 3 is used for acquiring a water surface image containing floating garbage;
the control panel is used for inputting the water surface image into the trained improved yolov5 model for identification, and the type and the position of the floating garbage are obtained.
In one embodiment, the upper computer is further configured to modify the yolov5 model to obtain a modified yolov5 model, and to train the modified yolov5 model to obtain a trained modified yolov5 model.
Optionally, the unmanned cleaning boat further comprises a baffle 6, the baffle 6 being adapted to concentrate the waste on a waste sorting deck 7.
Optionally, the rear boat 2 comprises a detection module, a garbage recycling bin 8 and a sending module;
the garbage collection bin 8 comprises a plurality of bins, and different bins are used for storing different types of floating garbage;
the detection module is used for detecting whether the stored garbage amount in the rear boat 2 reaches a warning line or not;
the sending module is used for sending GPS positioning to a standby boat which does not store floating garbage when the stored garbage amount in the rear boat 2 reaches a warning line; the standby boat is used for going to the position corresponding to the GPS positioning after the GPS positioning is received;
after the spare boat reaches the position corresponding to the GPS positioning, the upper computer is used for controlling the separation of the front boat 1 and the rear boat 2;
the upper computer is also used for controlling the connection of the front boat 1 and the standby boat to form a new unmanned cleaning boat;
the upper computer is also used for controlling the rear boat 2 to return to the shore.
Alternatively, the structure of the spare boat is identical to that of the rear boat 2.
According to the unmanned cleaning boat control method and the unmanned cleaning boat control system, the separable unmanned cleaning boat structure is designed, the spare rear boat 2 can be replaced after garbage is fully loaded, and the time required for returning to and fro for warehouse cleaning is shortened. Compared with the traditional unmanned cleaning boat, the method is more flexible in work, can effectively improve the working efficiency and expands the working range.
While preferred embodiments of the present invention have been described, additional variations and modifications in those embodiments may occur to those skilled in the art once they learn of the basic inventive concepts. It is therefore intended that the following claims be interpreted as including the preferred embodiments and all such alterations and modifications as fall within the scope of the invention.
It will be apparent to those skilled in the art that various modifications and variations can be made to the present invention without departing from the spirit or scope of the invention. Thus, it is intended that the present invention also include such modifications and alterations insofar as they come within the scope of the appended claims or the equivalents thereof.

Claims (10)

1. A method for sorting and locating refuse, comprising:
acquiring a water surface image containing floating garbage;
inputting the water surface image into a trained improved yolov5 model for recognition to obtain the type and position of the floating garbage,
wherein the improved yolov5 model comprises a backbond part, the backbond part comprises a Resnet network, a Batchnormalization layer and a dropout layer are added into the Resnet network, a CBAM module is inserted after each block of the Resnet network,
wherein, add the batch normalization layer and the dropout layer in the Resnet network, include:
a batch normalization layer is added after the convolutional layer of each residual block and identity block, and a dropout layer is added before the last full-join layer.
2. A method of sorting and locating refuse according to claim 1, characterized in that before acquiring the image of the water surface containing floating refuse, it further comprises:
improving the yolov5 model to obtain an improved yolov5 model;
and training the improved yolov5 model to obtain the trained improved yolov5 model.
3. The method of claim 1, wherein the modified yolov5 model further comprises a neg component and a Head component.
4. The garbage classification and localization method of claim 2, wherein training the modified yolov5 model comprises:
setting classification types of the water surface garbage;
acquiring a data set for training;
performing enhancement processing on the data set to obtain an enhanced data set;
dividing the enhanced data set into a training set, a verification set and a test set according to a set proportion;
performing label coding on the training set, the verification set and the test set;
selecting a loss function, an activation function, an optimizer, an evaluation function and a batch size respectively;
and training the improved yolov5 model by using the training set, the verification set and the test set after label coding to obtain the trained improved yolov5 model.
5. A method of unmanned cleaning boat control, comprising:
obtaining the type and location of floating refuse according to the refuse classification and positioning method of any one of claims 1-4;
controlling the unmanned cleaning boat to go to the position, wherein the unmanned cleaning boat comprises a front boat and a rear boat, and the front boat and the rear boat are designed in a separable way;
controlling the unmanned cleaning boat to salvage the floating garbage;
and conveying the floating garbage to the position corresponding to the rear boat according to the type of the floating garbage.
6. The method of claim 5, further comprising, after transferring the floating garbage to the location corresponding to the rear boat:
detecting whether the stored garbage amount in the rear boat reaches a warning line, if so, sending GPS positioning to a standby boat which does not store floating garbage; the standby boat is used for going to the position corresponding to the GPS positioning after the GPS positioning is received;
after the spare boat reaches the position corresponding to the GPS positioning, controlling the separation of the front boat and the rear boat;
controlling the front boat and the standby boat to be connected to form a new unmanned cleaning boat;
and controlling the rear boat to return to the shore.
7. The unmanned cleaning boat control system is characterized by comprising an upper computer and an unmanned cleaning boat, wherein the unmanned cleaning boat comprises a front boat and a rear boat which are designed in a separable way;
the unmanned cleaning boat further comprises a camera, a control panel, a garbage collection module, a conveyor belt and a garbage classification platform;
the camera, the control board and the upper computer are mutually matched, and the type and the position of the floating garbage are obtained according to the garbage classification and positioning method of any one of claims 1-4;
the control panel is also used for controlling the unmanned cleaning boat to go to the position;
the control panel is also used for controlling the garbage collection module to salvage the floating garbage;
the control panel is also used for controlling the conveyor belt to transmit the floating garbage to the garbage classification platform;
the control panel is also used for controlling the garbage classification platform to transmit the floating garbage to the position corresponding to the rear boat according to the type of the floating garbage.
8. The unmanned cleaning boat control system of claim 7, wherein the unmanned cleaning boat further comprises a baffle for focusing the waste on the waste sorting deck.
9. The unmanned cleaning boat control system of claim 7, wherein the rear boat comprises a detection module, a garbage collection bin, and a transmission module;
the garbage recycling bin comprises a plurality of bins, and different bins are used for storing different types of floating garbage;
the detection module is used for detecting whether the stored garbage amount in the rear boat reaches a warning line or not;
the sending module is used for sending GPS positioning to the spare boat which does not store floating garbage when the stored garbage amount in the rear boat reaches a warning line; the standby boat is used for going to the position corresponding to the GPS positioning after the GPS positioning is received;
after the spare boat reaches the position corresponding to the GPS positioning, the upper computer is used for controlling the separation of the front boat and the rear boat;
the upper computer is also used for controlling the connection of the front boat and the standby boat to form a new unmanned cleaning boat;
the upper computer is also used for controlling the back boat to return to the shore.
10. An unmanned cleaning boat control system according to claim 9, wherein the configuration of the spare boat is identical to the configuration of the rear boat.
CN202310515917.4A 2023-05-09 2023-05-09 Garbage classification and positioning method, unmanned cleaning boat control method and system Pending CN116310597A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202310515917.4A CN116310597A (en) 2023-05-09 2023-05-09 Garbage classification and positioning method, unmanned cleaning boat control method and system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202310515917.4A CN116310597A (en) 2023-05-09 2023-05-09 Garbage classification and positioning method, unmanned cleaning boat control method and system

Publications (1)

Publication Number Publication Date
CN116310597A true CN116310597A (en) 2023-06-23

Family

ID=86824285

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202310515917.4A Pending CN116310597A (en) 2023-05-09 2023-05-09 Garbage classification and positioning method, unmanned cleaning boat control method and system

Country Status (1)

Country Link
CN (1) CN116310597A (en)

Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2016023080A1 (en) * 2014-08-12 2016-02-18 Frozen Motion Pte Ltd Marine craft for performing surface operations
CN111469131A (en) * 2020-05-10 2020-07-31 上海大学 Unmanned ship water surface garbage cleaning control system and method with mechanical arm
CN112215131A (en) * 2020-10-10 2021-01-12 李睿宸 Automatic garbage picking system and manual operation and automatic picking method thereof
CN112488162A (en) * 2020-11-17 2021-03-12 中南民族大学 Garbage classification method based on active learning
CN113049084A (en) * 2021-03-16 2021-06-29 电子科技大学 Attention mechanism-based Resnet distributed optical fiber sensing signal identification method
CN113850242A (en) * 2021-11-30 2021-12-28 北京中超伟业信息安全技术股份有限公司 Storage abnormal target detection method and system based on deep learning algorithm
CN113989749A (en) * 2021-11-17 2022-01-28 交控科技股份有限公司 Station passenger flow monitoring method, electronic device and storage medium
CN115027630A (en) * 2022-07-21 2022-09-09 张恺知 Unmanned ship on water
CN217674749U (en) * 2022-05-16 2022-10-28 广东工业大学 Garbage classification recycling vehicle with visual identification function
CN115276855A (en) * 2022-06-16 2022-11-01 宁波大学 ResNet-CBAM-based spectrum sensing method

Patent Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2016023080A1 (en) * 2014-08-12 2016-02-18 Frozen Motion Pte Ltd Marine craft for performing surface operations
CN111469131A (en) * 2020-05-10 2020-07-31 上海大学 Unmanned ship water surface garbage cleaning control system and method with mechanical arm
CN112215131A (en) * 2020-10-10 2021-01-12 李睿宸 Automatic garbage picking system and manual operation and automatic picking method thereof
CN112488162A (en) * 2020-11-17 2021-03-12 中南民族大学 Garbage classification method based on active learning
CN113049084A (en) * 2021-03-16 2021-06-29 电子科技大学 Attention mechanism-based Resnet distributed optical fiber sensing signal identification method
CN113989749A (en) * 2021-11-17 2022-01-28 交控科技股份有限公司 Station passenger flow monitoring method, electronic device and storage medium
CN113850242A (en) * 2021-11-30 2021-12-28 北京中超伟业信息安全技术股份有限公司 Storage abnormal target detection method and system based on deep learning algorithm
CN217674749U (en) * 2022-05-16 2022-10-28 广东工业大学 Garbage classification recycling vehicle with visual identification function
CN115276855A (en) * 2022-06-16 2022-11-01 宁波大学 ResNet-CBAM-based spectrum sensing method
CN115027630A (en) * 2022-07-21 2022-09-09 张恺知 Unmanned ship on water

Similar Documents

Publication Publication Date Title
CN107617573B (en) Logistics code identification and sorting method based on multitask deep learning
CN107368614B (en) Image retrieval method and device based on deep learning
Demir et al. Detection of land-cover transitions in multitemporal remote sensing images with active-learning-based compound classification
EP3327583B1 (en) Method and device for searching a target in an image
Oquab et al. Weakly supervised object recognition with convolutional neural networks
Wang et al. Deep proposal and detection networks for road damage detection and classification
CN114241260B (en) Open set target detection and identification method based on deep neural network
CN110647907B (en) Multi-label image classification algorithm using multi-layer classification and dictionary learning
Haloi A novel plsa based traffic signs classification system
Sharmanska et al. Ambiguity helps: Classification with disagreements in crowdsourced annotations
CN112733936A (en) Recyclable garbage classification method based on image recognition
CN110414626A (en) A kind of pig variety ecotype method, apparatus and computer readable storage medium
CN110674844A (en) Intelligent container increment learning training method
CN115187870A (en) Marine plastic waste material identification method and system, electronic equipment and storage medium
CN116310597A (en) Garbage classification and positioning method, unmanned cleaning boat control method and system
Chaurasia et al. Recognition of handwritten numerals of various indian regional languages using deep learning
CN117373111A (en) AutoHOINet-based human-object interaction detection method
CN112784788B (en) Low-resolution target detection method
CN102169493A (en) Method for automatically identifying experimental scheme from literatures
WO2022185340A1 (en) Material detector
CN113344102A (en) Target image identification method based on image HOG characteristics and ELM model
Lee et al. Integrated segmentation and recognition of connected handwritten characters with recurrent neural network
Pu Application of active learning algorithm in handwriting recognition numbers
Wang et al. Analysis of Influencing Factors of Water Traffic Accidents Severity Based on LIBSVM
Addagarla et al. Single Stage Deep Transfer Learning Model for Apparel Detection and Classification for E-Commerce

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
RJ01 Rejection of invention patent application after publication
RJ01 Rejection of invention patent application after publication

Application publication date: 20230623