CN113158965A - Vision-imitated recognition method, device and medium for realizing garbage recognition of sea drift - Google Patents
Vision-imitated recognition method, device and medium for realizing garbage recognition of sea drift Download PDFInfo
- Publication number
- CN113158965A CN113158965A CN202110500850.8A CN202110500850A CN113158965A CN 113158965 A CN113158965 A CN 113158965A CN 202110500850 A CN202110500850 A CN 202110500850A CN 113158965 A CN113158965 A CN 113158965A
- Authority
- CN
- China
- Prior art keywords
- sea
- garbage
- image data
- background
- time
- 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.)
- Granted
Links
- 239000010813 municipal solid waste Substances 0.000 title claims abstract description 100
- 238000000034 method Methods 0.000 title claims abstract description 30
- 238000012545 processing Methods 0.000 claims abstract description 25
- 238000012549 training Methods 0.000 claims abstract description 19
- 238000012544 monitoring process Methods 0.000 claims abstract description 17
- 238000001514 detection method Methods 0.000 claims abstract description 5
- 230000000007 visual effect Effects 0.000 claims description 9
- 238000004590 computer program Methods 0.000 claims description 6
- 238000001914 filtration Methods 0.000 claims description 6
- 238000007781 pre-processing Methods 0.000 claims description 3
- 239000004576 sand Substances 0.000 description 2
- 239000002699 waste material Substances 0.000 description 2
- 208000034699 Vitreous floaters Diseases 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 210000004556 brain Anatomy 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 239000006260 foam Substances 0.000 description 1
- 238000002372 labelling Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000000630 rising effect Effects 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/194—Segmentation; Edge detection involving foreground-background segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- 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
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Multimedia (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Image Processing (AREA)
Abstract
The invention relates to a vision-imitated recognition method for realizing garbage recognition of a sea float, which comprises the following steps of: dividing a monitoring sea area into a plurality of areas, and setting up image acquisition equipment in the monitoring sea area to acquire image data of each area; selecting background images corresponding to the areas and placing the background images into a background database; training a target detection model through a picture of the garbage from the sea drift to generate a garbage from the sea drift recognition model; acquiring real-time image data of each area through image acquisition equipment, comparing the real-time image data of each area with a corresponding background image, performing XOR processing according to a comparison result, eliminating the same part of the real-time image data and the background image, and keeping the part of the real-time image data, which is different from the background image, as a target to be identified; and identifying the target to be identified through the garbage from sea floating identification model, outputting the target to be identified which is identified as garbage from sea floating, and performing iterative training on the garbage from sea floating identification model after marking the target to be identified which cannot be identified.
Description
Technical Field
The invention relates to a vision-imitated recognition method for realizing garbage floating on the sea, belonging to the technical field of environmental management and image recognition.
Background
On a global scale, the garbage from sea floating is a serious problem, and the situation is particularly severe in Asia. More than 80% of the world's marine floating waste is reported to come from asia. Most of the garbage floating on the sea is difficult to naturally degrade and can only be collected and cleaned by manpower. And a large amount of garbage on the sea drift is far away from the land, so that the garbage needs to be collected from the open sea and transported back to the land for treatment. A large amount of manpower and material resources are consumed in the whole process, and a large burden is brought to local finance.
In the prior art, the method comprises the steps of monitoring the floating garbage on the sea surface by a video monitoring method, manually checking foreign matters appearing in a video, determining whether the foreign matters are the floating garbage, and informing sea surface governing personnel to clear the floating garbage after determining that the foreign matters are the garbage; the manual monitoring is low in efficiency, workers on duty need to check foreign matters in the video all the time, and the possibility of wrong identification exists in manual identification, so that fishes or fishing tools and the like can be identified as floating garbage, and wrong information is notified to sea surface management personnel, and resource waste is caused.
Disclosure of Invention
In order to solve the problems in the prior art, the invention provides a vision-imitated identification method for realizing the identification of the garbage floating on the sea.
The technical scheme of the invention is as follows:
the first technical scheme is as follows:
a vision-imitated recognition method for realizing garbage recognition of sea floaters comprises the following steps:
determining a monitoring sea area; dividing a monitoring sea area into a plurality of areas, and setting at least one image acquisition device in the monitoring sea area to acquire image data of each area;
establishing a background database; selecting background images which correspond to the areas one by one according to the image data of the areas, and putting the background images into a background database, wherein the background images are images of the sea surface which do not contain foreign matters;
training a model; collecting a plurality of images of the garbage from the sea-floating as a training sample set, and training a target detection model through the training sample set to generate a garbage from the sea-floating recognition model;
performing vision imitation treatment; acquiring real-time image data of each area through image acquisition equipment, comparing the real-time image data of each area with a corresponding background image, performing XOR processing according to a comparison result, eliminating the same part of the real-time image data and the background image, and keeping the part of the real-time image data, which is different from the background image, as a target to be identified;
identifying the garbage floating on the sea; and identifying the target to be identified through the garbage from sea floating identification model, outputting the target to be identified as garbage from sea floating, and performing iterative training on the garbage from sea floating identification model after marking the target to be identified which cannot be identified.
Further, the method for selecting the background image specifically comprises the following steps:
firstly, manually selecting background images of all areas, putting the background images into a background database, and recording selection time;
setting a time threshold T0, if the time interval between the current time and the selection time is equal to the time threshold T0, automatically acquiring real-time image data of each area through image acquisition equipment, and identifying whether the real-time image data contains the garbage from the sea floating through the garbage identification model;
if the real-time image data of an area does not contain the garbage from the sea drift, updating the background image of the area in the background database into the current image data and recording the updating time;
if the real-time image data of an area contains the garbage from sea floating, continuously monitoring the real-time image data of the area through a garbage from sea floating recognition model until the real-time image data of the area does not contain the garbage from sea floating, updating the background image of the area in a background database into current image data and recording the updating time;
the background images of the respective regions are continuously updated in accordance with the update time and the time threshold T0 in the above-described procedure.
Further, in the step of processing the simulated vision, the method further comprises a step of preprocessing an image, specifically:
respectively determining the sea surface contour of each region according to the image data of each region;
respectively making and storing cutting templates of each region according to the sea surface contour of each region;
and respectively carrying out image cutting on the real-time image data and the background image of each region through the cutting template of each region, and removing parts except the sea surface outline in the real-time image data and the background image.
Further, the specific steps of the xor processing are as follows:
carrying out gray level processing on the background image to obtain a background gray level image;
after the background gray image is subjected to smooth filtering processing, acquiring a gray value corresponding to each pixel point in the background gray image;
performing the gray processing and the smooth filtering processing on the real-time image data to obtain a gray value corresponding to each pixel point in a real-time gray image;
setting a gray difference threshold T1, comparing the gray value of each pixel point in the real-time gray image with the gray value of each pixel point in the background gray image one by one, acquiring a gray difference layout image with pixel points corresponding to gray difference values one by one, binarizing the gray difference layout image through a gray difference threshold T1, setting the pixel points with the gray difference values larger than the gray difference threshold T1 as black points, and setting the pixel points with the gray difference values smaller than the gray difference threshold T1 as white points, and generating a binary image;
and determining the contour of the target to be identified according to the binarized black point contour, and extracting the target to be identified from the real-time image data according to the contour of the target to be identified.
The second technical scheme is as follows:
an imitative visual recognition device for realizing garbage collection from sea drift comprises a memory, a processor and a computer program which is stored on the memory and can run on the processor, wherein the processor executes the program to realize the method of the first technical scheme.
The third technical scheme is as follows:
a visual imitation recognition medium for realizing the recognition of garbage from sea drift, on which a computer program is stored, characterized in that the program, when being executed by a processor, realizes the method according to the first technical aspect.
The invention has the following beneficial effects:
1. according to the vision-imitated identification method for realizing the identification of the garbage floating on the sea, the garbage floating on the sea can be automatically identified by training the garbage floating on the sea identification model, the target to be identified, which is different from the background image, on the sea surface is segmented by vision-imitated processing, and the target to be identified is identified by the garbage floating on the sea identification model, so that the identification efficiency of the garbage floating on the sea is improved, the workload of workers is greatly reduced, and the probability of the garbage floating on the sea identification error is reduced.
2. The invention relates to an imitation visual identification method for realizing the identification of garbage on the sea drift, which reduces the interference of background noise by continuously updating a background image and improves the identification accuracy of a target to be extracted.
3. The invention relates to a vision-imitated identification method for realizing garbage identification of sea floating, which cuts image data by setting a template, can eliminate interference outside the sea level in the image data, for example, the image data comprises part of sand beach, sky or reef, and the like, and can eliminate the interference to avoid causing the interference.
Drawings
FIG. 1 is a flow chart of an embodiment of the present invention;
FIG. 2 is an exemplary diagram of performing XOR processing in an embodiment of the present invention.
Detailed Description
The invention is described in detail below with reference to the figures and the specific embodiments.
The first embodiment is as follows:
referring to fig. 1, a visual imitation identification method for realizing the identification of garbage on the sea drift comprises the following steps:
determining a monitoring sea area; dividing a monitoring sea area into a plurality of areas, and setting at least one image acquisition device in the monitoring sea area to acquire image data of each area, wherein the image data can be video streams or pictures;
establishing a background database; selecting background images which correspond to the areas one by one according to the image data of the areas, and putting the background images into a background database, wherein the background images are images of the sea surface which do not contain foreign matters;
training a model; collect the picture of a plurality of sea floating garbage, carry out artificial mark to the sea floating garbage in the picture, mark rubbish classification, rubbish classification can be "belong to the sea floating garbage" or "do not belong to the big classification such as sea floating garbage", can also subdivide big classification, for example under the classification of "belong to the sea floating garbage", can specifically subdivide rubbish type, if: branches, water bottles, trash bags, foam, and the like; under the classification "not belonging to the garbage on sea", the object categories can be subdivided specifically, for example: marine organisms, reef, etc. Taking the artificially labeled picture as a training sample set, and training a target detection model through the training sample set to generate a garbage from sea drift recognition model; when the target detection model is trained, the sample and the approximate contour of the currently labeled target are learned according to the garbage category information labeled in the picture of the marine floating garbage, then the trained marine floating garbage recognition model is used for analyzing the image data, and the corresponding target in the image data can be quickly found according to the learned parameters and framed.
Performing vision imitation treatment; referring to fig. 2, the real-time image data of each region is acquired by the image acquisition device, the real-time image data of each region is compared with the corresponding background image, and xor processing is performed according to the comparison result, where the upper and lower left pictures in fig. 2 are the background image of a region and the real-time image data of the region. Eliminating the same part of the real-time image data and the background image, reserving the part of the real-time image data different from the background image as a target to be identified, and setting the right picture in the figure 2 as the target to be identified; when the floating garbage is identified manually, firstly, a worker can memorize the picture of the sea surface, memorize the initial state of the sea surface, and when a foreign matter appears on the sea surface, the worker can inform the brain of the foreign matter appearing when observing that the current picture of the sea surface has a part different from the memorized picture.
Identifying the garbage floating on the sea; identifying the target to be identified through a garbage on sea surface identification model, outputting the target to be identified as garbage on sea surface, simultaneously outputting alarm information to an upper management system to inform a sea surface treating person of the existence of garbage in the area, and also transmitting an identified garbage on sea surface image to the sea surface treating person through the upper management system to inform the sea surface treating person of the shape profile and the position of the garbage on sea surface; for the target to be identified which is output as the garbage which does not belong to the sea floating, such as reef, marine life and the like, no treatment is carried out; and sending the target to be identified which cannot be identified into a database to be labeled, wherein a worker can browse and label the database to be labeled by the upper system software, manually label garbage categories on the target to be identified which cannot be identified, and then sending the image data of the target to be identified after labeling into a garbage from sea level to carry out iterative training.
The rubbish is floated to the sea in this embodiment through training sea, can the automatic identification image data in the rubbish is floated to the sea, handles through imitative vision, cuts apart out the target of treating discernment different with the background image on the sea, and the discernment target is treated to the rubbish is floated to the rethread sea and the rubbish is discerned to the rubbish recognition model, improves the discernment efficiency that the rubbish was floated to the sea, and the staff's that significantly reduces work load reduces the probability that rubbish was floated to the sea discernment mistake.
Example two:
in this embodiment, the method for selecting the background image specifically includes:
firstly, manually selecting background images of each area, putting the background images into a background database, and recording selection time, such as 20 minutes and 15 seconds when recording 1 month, 31 days and 5 days of 2020;
setting a time threshold T0, for example, ten minutes, if the time interval between the current time and the selection time is equal to the time threshold T0, namely the time interval reaches 1 month, 31 days, 5 days, 30 minutes and 15 seconds in 2020, automatically acquiring real-time image data of each area through image acquisition equipment, and identifying whether the real-time image data contains the garbage from the sea floating through the garbage identification model;
if the real-time image data of an area does not contain the garbage from the sea drift, updating the background image of the area in the background database into the current image data and recording the updating time;
if the real-time image data of an area contains the garbage from sea floating, continuously monitoring the real-time image data of the area through a garbage from sea floating recognition model until the real-time image data of the area does not contain the garbage from sea floating, updating the background image of the area in a background database into current image data and recording the updating time;
continuously updating the background images of the areas according to the updating time and the time threshold T0 according to the steps; in the embodiment, the background image is continuously updated, so that the background noise between the background image and the real-time image data is minimum when the exclusive-or processing is performed, for example, the background noise is color change of the sea surface caused by sunlight or sea surface waves caused by tide rising and falling, the interference of the background noise is reduced by continuously updating the background image, and the identification accuracy of the target to be extracted is improved.
In this embodiment, the step of processing the pseudo-visual image further includes a step of preprocessing an image, specifically:
respectively determining the sea surface contour of each region according to the image data of each region;
respectively making and storing cutting templates of each region according to the sea surface contour of each region;
and respectively carrying out image cutting on the real-time image data and the background image of each region through the cutting template of each region, and removing parts except the sea surface outline in the real-time image data and the background image. By the scheme of the embodiment, the interference outside the sea surface in the image data can be eliminated, for example, the image data comprises part of sand beach, sky or reef, and the like, so that the interference can be avoided.
In this embodiment, the specific steps of the xor processing are as follows:
carrying out gray level processing on the background image to obtain a background gray level image;
after the background gray image is subjected to smooth filtering processing, acquiring a gray value corresponding to each pixel point in the background gray image; in this embodiment, a horizontal rectangular coordinate system is made with a pixel point in the background gray-scale image as an origin, the coordinate of the pixel point is (x, y), the gray value corresponding to the pixel point is G (x, y), which indicates that the gray value of the pixel point with the abscissa as x and the ordinate as y is G.
Performing the gray processing and the smooth filtering processing on the real-time image data to obtain a gray value corresponding to each pixel point in a real-time gray image, and recording the gray value corresponding to each pixel point in the real-time gray image as K (x, y);
setting a gray difference threshold T1, for example, T1 is 50, the gray threshold T1 is set according to human experience, and a value with a better effect can be selected through multiple experiments, comparing the gray value of each pixel in the real-time gray map with the gray value of each pixel in the background gray map one by one, obtaining a gray difference layout with one-to-one correspondence between the pixel point and the gray difference value, for example, G (1,2) is 20, K (1,2) is 25, the gray difference of the pixel point (1,2) is 5, G (10,20) is 1, K (10,20) is 255, the gray difference of the pixel point (10,20) is 254, binarizing the gray difference layout through the gray difference threshold T1, setting the pixel point with the gray difference value larger than the gray difference threshold T1 as a black point, the pixel point with the gray difference value smaller than the gray difference threshold T1 as a white point, generating a binarized image, that is, the point (1,2) is set as a white point, the dots (10,20) are set to black dots; and setting a gray threshold value to quickly identify the shape contour of the target to be identified in a gray contrast mode because the pixel of the target to be identified which is inconsistent with the background sea surface is inconsistent with the pixel of the background sea surface when the target to be identified which is inconsistent with the background image appears.
And determining the contour of the target to be identified according to the binarized black point contour, and extracting the target to be identified from the real-time image data according to the contour of the target to be identified.
Example three:
an imitative visual identification device for realizing garbage identification of a floating sea comprises a memory, a processor and a computer program which is stored on the memory and can run on the processor, wherein the processor executes the program to realize the method according to any embodiment of the invention.
Example four:
a visual imitation recognition medium for realizing garbage from sea drift, on which a computer program is stored, characterized in that the program, when being executed by a processor, realizes the method according to any one of the embodiments of the present invention.
The above description is only an embodiment of the present invention, and not intended to limit the scope of the present invention, and all modifications of equivalent structures and equivalent processes performed by the present specification and drawings, or directly or indirectly applied to other related technical fields, are included in the scope of the present invention.
Claims (6)
1. A vision-imitated recognition method for realizing garbage recognition of a sea float is characterized by comprising the following steps:
determining a monitoring sea area; dividing a monitoring sea area into a plurality of areas, and setting at least one image acquisition device in the monitoring sea area to acquire image data of each area;
establishing a background database; selecting background images which correspond to the areas one by one according to the image data of the areas, and putting the background images into a background database, wherein the background images are images of the sea surface which do not contain foreign matters;
training a model; collecting a plurality of images of the garbage from the sea-floating as a training sample set, and training a target detection model through the training sample set to generate a garbage from the sea-floating recognition model;
performing vision imitation treatment; acquiring real-time image data of each area through image acquisition equipment, comparing the real-time image data of each area with a corresponding background image, performing XOR processing according to a comparison result, eliminating the same part of the real-time image data and the background image, and keeping the part of the real-time image data, which is different from the background image, as a target to be identified;
identifying the garbage floating on the sea; and identifying the target to be identified through the garbage from sea floating identification model, outputting the target to be identified as garbage from sea floating, and performing iterative training on the garbage from sea floating identification model after marking the target to be identified which cannot be identified.
2. The vision-imitated recognition method for realizing garbage floating on sea of claim 1, wherein the method for selecting the background image specifically comprises the following steps:
firstly, manually selecting background images of all areas, putting the background images into a background database, and recording selection time;
setting a time threshold T0, if the time interval between the current time and the selection time is equal to the time threshold T0, automatically acquiring real-time image data of each area through image acquisition equipment, and identifying whether the real-time image data contains the garbage from the sea floating through the garbage identification model;
if the real-time image data of an area does not contain the garbage from the sea drift, updating the background image of the area in the background database into the current image data and recording the updating time;
if the real-time image data of an area contains the garbage from sea floating, continuously monitoring the real-time image data of the area through a garbage from sea floating recognition model until the real-time image data of the area does not contain the garbage from sea floating, updating the background image of the area in a background database into current image data and recording the updating time;
the background images of the respective regions are continuously updated in accordance with the update time and the time threshold T0 in the above-described procedure.
3. The visual imitation identification method for realizing the identification of the garbage from the sea drift according to claim 1, further comprising an image preprocessing step in the visual imitation processing step, specifically:
respectively determining the sea surface contour of each region according to the image data of each region;
respectively making and storing cutting templates of each region according to the sea surface contour of each region;
and respectively carrying out image cutting on the real-time image data and the background image of each region through the cutting template of each region, and removing parts except the sea surface outline in the real-time image data and the background image.
4. The vision-imitated recognition method for realizing garbage floating on sea of claim 1, wherein the specific steps of the exclusive or processing are as follows:
carrying out gray level processing on the background image to obtain a background gray level image;
after the background gray image is subjected to smooth filtering processing, acquiring a gray value corresponding to each pixel point in the background gray image;
performing the gray processing and the smooth filtering processing on the real-time image data to obtain a gray value corresponding to each pixel point in a real-time gray image;
setting a gray difference threshold T1, comparing the gray value of each pixel point in the real-time gray image with the gray value of each pixel point in the background gray image one by one, acquiring a gray difference layout image with pixel points corresponding to gray difference values one by one, binarizing the gray difference layout image through a gray difference threshold T1, setting the pixel points with the gray difference values larger than the gray difference threshold T1 as black points, and setting the pixel points with the gray difference values smaller than the gray difference threshold T1 as white points, and generating a binary image;
and determining the contour of the target to be identified according to the binarized black point contour, and extracting the target to be identified from the real-time image data according to the contour of the target to be identified.
5. An imitation vision recognition device for realizing garbage from sea drift recognition, comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the processor implements the method according to any one of claims 1 to 4 when executing the program.
6. A visual imitation identification medium for realizing the identification of garbage from sea drifts, on which a computer program is stored, characterized in that the program, when being executed by a processor, realizes the method according to any one of claims 1 to 4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110500850.8A CN113158965B (en) | 2021-05-08 | 2021-05-08 | Imitative visual recognition method, device and medium for realizing sea-bleaching garbage recognition |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110500850.8A CN113158965B (en) | 2021-05-08 | 2021-05-08 | Imitative visual recognition method, device and medium for realizing sea-bleaching garbage recognition |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113158965A true CN113158965A (en) | 2021-07-23 |
| CN113158965B CN113158965B (en) | 2024-03-19 |
Family
ID=76874028
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110500850.8A Active CN113158965B (en) | 2021-05-08 | 2021-05-08 | Imitative visual recognition method, device and medium for realizing sea-bleaching garbage recognition |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113158965B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114549826A (en) * | 2022-02-25 | 2022-05-27 | 广州大学 | Energy-saving marine garbage collection return algorithm and system based on machine vision |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20190065859A1 (en) * | 2016-12-09 | 2019-02-28 | Hitachi Kokusai Electric Inc. | Marine intrusion detection system and method |
| WO2019207577A1 (en) * | 2018-04-25 | 2019-10-31 | Imagry (Israel) Ltd. | System and method for training a machine-learning model to identify real-world elements |
| CN110472628A (en) * | 2019-08-10 | 2019-11-19 | 四创科技有限公司 | A kind of improvement Faster R-CNN network detection floating material method based on video features |
| CN111079724A (en) * | 2020-03-25 | 2020-04-28 | 速度时空信息科技股份有限公司 | Unmanned aerial vehicle-based sea floating garbage identification method |
| CN111259809A (en) * | 2020-01-17 | 2020-06-09 | 五邑大学 | Unmanned aerial vehicle coastline floating garbage inspection system based on DANet |
| CN111409780A (en) * | 2020-03-25 | 2020-07-14 | 国家海洋局北海海洋环境监测中心站 | Marine waste collection device |
-
2021
- 2021-05-08 CN CN202110500850.8A patent/CN113158965B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20190065859A1 (en) * | 2016-12-09 | 2019-02-28 | Hitachi Kokusai Electric Inc. | Marine intrusion detection system and method |
| WO2019207577A1 (en) * | 2018-04-25 | 2019-10-31 | Imagry (Israel) Ltd. | System and method for training a machine-learning model to identify real-world elements |
| CN110472628A (en) * | 2019-08-10 | 2019-11-19 | 四创科技有限公司 | A kind of improvement Faster R-CNN network detection floating material method based on video features |
| CN111259809A (en) * | 2020-01-17 | 2020-06-09 | 五邑大学 | Unmanned aerial vehicle coastline floating garbage inspection system based on DANet |
| CN111079724A (en) * | 2020-03-25 | 2020-04-28 | 速度时空信息科技股份有限公司 | Unmanned aerial vehicle-based sea floating garbage identification method |
| CN111409780A (en) * | 2020-03-25 | 2020-07-14 | 国家海洋局北海海洋环境监测中心站 | Marine waste collection device |
Non-Patent Citations (5)
| Title |
|---|
| BORGHGRAEF A等: "An evaluation of pixel-based methods for the detection of floating objects on the sea surface", 《EURASIP JOURNAL ON ADVANCES IN SIGNAL PROCESSING》, pages 451 - 461 * |
| DRISS MOUJAHID等: "Moving object detection zone using a block-based background model", 《HTTPS://DOI.ORG/10.1049/IET-CVI.2017.0136》, pages 1 - 9 * |
| 汤伟等: "改进卷积神经网络算法在水面漂浮垃圾检测中的应用", 《中国科技论文》, pages 1210 - 1216 * |
| 沈钰: "基于特征的无人机河湖图像智能处理系统", 《中国优秀硕士学位论文全文数据库》工程科技Ⅱ辑, pages 031 - 2 * |
| 田扬捷, 黄仁华, 杨虹, 周海燕, 李道棠: "海滨垃圾填埋场渗沥液对地下水系统的污染", 环境卫生工程, no. 01, pages 1 - 5 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114549826A (en) * | 2022-02-25 | 2022-05-27 | 广州大学 | Energy-saving marine garbage collection return algorithm and system based on machine vision |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113158965B (en) | 2024-03-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111986099B (en) | Tillage monitoring method and system based on convolutional neural network with residual error correction fused | |
| CN109376605B (en) | Electric power inspection image bird-stab-prevention fault detection method | |
| CN102867196B (en) | Based on the complicated sea remote sensing image Ship Detection of Gist feature learning | |
| CN111723657B (en) | River foreign matter detection method and device based on YOLOv3 and self-optimization | |
| CN112508079B (en) | Fine identification method, system, equipment, terminal and application of ocean frontal surface | |
| CN116994140A (en) | Cultivated land extraction method, device, equipment and medium based on remote sensing image | |
| CN111028255A (en) | Farmland area pre-screening method and device based on prior information and deep learning | |
| CN108564114B (en) | Human body fecal leucocyte automatic identification method based on machine learning | |
| CN111339902B (en) | Liquid crystal display indication recognition method and device for digital display instrument | |
| CN113191374B (en) | PolSAR image ridge line extraction method based on pyramid attention network | |
| CN106780504B (en) | It is a kind of based on distance keep level set beef image in carry on the back long flesh automatic division method | |
| CN111950812B (en) | Method and device for automatically identifying and predicting rainfall | |
| CN112733711B (en) | Remote sensing image damaged building extraction method based on multi-scale scene change detection | |
| CN114120094A (en) | Water pollution identification method and system based on artificial intelligence | |
| CN115497015A (en) | River floating pollutant identification method based on convolutional neural network | |
| CN110853041A (en) | Underwater pier component segmentation method based on deep learning and sonar imaging | |
| CN117036715A (en) | Deformation region boundary automatic extraction method based on convolutional neural network | |
| CN113158965B (en) | Imitative visual recognition method, device and medium for realizing sea-bleaching garbage recognition | |
| CN110991300A (en) | Automatic identification method for abnormal swelling state of dorking abdomen | |
| CN116433596A (en) | Slope vegetation coverage measuring method and device and related components | |
| CN115953678A (en) | Pavement damage detection method based on local gray extreme point and feature fusion | |
| Doxani et al. | Automatic change detection in urban areas under a scale-space, object-oriented classification framework | |
| CN115719456A (en) | Building fine identification method based on deep learning | |
| CN115147702A (en) | YOLACT + + based vehicle damage identification method, device, equipment and medium | |
| CN110570437A (en) | electric power channel automatic inspection data processing method based on boundary identification |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |