CN114005038A - Method for identifying illegal buildings near scenic spot - Google Patents
Method for identifying illegal buildings near scenic spot Download PDFInfo
- Publication number
- CN114005038A CN114005038A CN202111316369.XA CN202111316369A CN114005038A CN 114005038 A CN114005038 A CN 114005038A CN 202111316369 A CN202111316369 A CN 202111316369A CN 114005038 A CN114005038 A CN 114005038A
- Authority
- CN
- China
- Prior art keywords
- image
- road
- dbl
- violation building
- neural network
- 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 31
- 238000003062 neural network model Methods 0.000 claims abstract description 21
- 230000011218 segmentation Effects 0.000 claims abstract description 11
- 238000012549 training Methods 0.000 claims abstract description 7
- 238000004364 calculation method Methods 0.000 claims abstract description 4
- 101000615238 Homo sapiens Proto-oncogene DBL Proteins 0.000 claims description 48
- 239000013598 vector Substances 0.000 claims description 29
- 238000001514 detection method Methods 0.000 claims description 21
- 238000011176 pooling Methods 0.000 claims description 15
- 238000013527 convolutional neural network Methods 0.000 claims description 9
- 101100194362 Schizosaccharomyces pombe (strain 972 / ATCC 24843) res1 gene Proteins 0.000 claims description 6
- 101100194363 Schizosaccharomyces pombe (strain 972 / ATCC 24843) res2 gene Proteins 0.000 claims description 6
- 230000004927 fusion Effects 0.000 claims description 6
- 238000012545 processing Methods 0.000 claims description 6
- 230000004913 activation Effects 0.000 claims description 3
- 230000008569 process Effects 0.000 claims description 3
- 238000005070 sampling Methods 0.000 claims description 3
- 239000013589 supplement Substances 0.000 claims description 3
- 238000012790 confirmation Methods 0.000 abstract description 4
- 230000008859 change Effects 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
Images
Classifications
-
- 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
- G06N3/045—Combinations of networks
-
- 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
Abstract
The invention relates to a method for identifying illegal buildings near scenic spots, which comprises the following steps: constructing a deep neural network model, and training the constructed deep neural network model; outputting the detected position area of each suspected violation building and the road segmentation mask of the road section to be detected of the target road by the trained deep neural network model; calculating the area of each suspected violation building and the distance between each suspected violation building and the road section to be detected of the target road; and presetting a distance discrimination threshold, and if the distance calculation value between the suspected violation building and the road section to be detected of the target road is smaller than the preset distance discrimination threshold for each suspected violation building, judging the suspected violation building as the violation building. The invention has the advantages of automatically identifying the illegal buildings, improving the identification accuracy of the illegal buildings, reducing the participation of manpower and accelerating the identification efficiency so as to facilitate the on-site confirmation and supervision of patrolmen.
Description
Technical Field
The invention relates to a method for identifying illegal buildings near scenic spots.
Background
The traditional monitoring and management of illegal buildings near the scenic spot mainly depends on the report, discovery and confirmation of the masses, and the illegal buildings are forbidden and cannot be discovered in time, so that the development and construction of the urban scenic spot are hindered. Some methods identify the roadside illegal buildings by analyzing the road images, but the existing identification methods compare the acquired images with the original historical images one by one, and the identification methods need to repeatedly call the historical images, mostly need to combine manpower for identification and detection, consume resources and reduce efficiency.
Disclosure of Invention
The invention aims to provide a method for identifying illegal buildings near scenic spots, which aims to solve the problems in the background technology.
In order to achieve the purpose, the invention provides the following technical scheme:
a method for identifying illegal buildings near a scenic spot comprises the following steps:
constructing a deep neural network model, and training the constructed deep neural network model until the deep neural network model is trained to preset target parameters;
outputting the detected position area of each suspected violation building and the road segmentation mask of the road section to be detected of the target road by the trained deep neural network model; calculating the area of each suspected violation building and the distance between each suspected violation building and the road section to be detected of the target road; and presetting a distance discrimination threshold, and if the distance calculation value between the suspected violation building and the road section to be detected of the target road is smaller than the preset distance discrimination threshold for each suspected violation building, judging the suspected violation building as the violation building.
As a further scheme of the invention: the constructed deep neural network model comprises three modules, namely a road segmentation module, a violation building detection module and a fusion ranging module.
As a further scheme of the invention: the method for processing the original image by the road segmentation module comprises the following steps:
carrying out first vector convolution operation on the original image, and reducing the original image into 1/2 after pooling;
carrying out second vector convolution operation on the image, and reducing the image into 1/4 after pooling;
performing vector convolution operation and pooling for the third time on the image, reducing the image into 1/8 parts of the original image, and keeping the pooled feature map;
performing fourth vector convolution operation and pooling on the image, reducing the image to 1/16 of the original image, and keeping a pooled feature map;
and performing fifth vector convolution operation and pooling on the image, reducing the image into 1/32 of the original image, changing the full connection in the original CNN convolutional neural network operation into sixth vector convolution operation and seventh vector convolution operation, changing the number of feature maps of the image into 1/32 of the original image, obtaining an image heat map with a 1/32-size heat map, a 1/16-size feature map and a 1/8-size feature map, performing an up-sampling process on the 1/32-size heat map, and performing deconvolution on the image from the fifth vector convolution operation to the first vector convolution operation to supplement details.
As a further scheme of the invention: the method for processing the image by the illegal building detection module comprises the following steps:
using yolov3 object detection technical feature extractor Darknet-53, wherein the main structures of 53 convolutional layers are DBL, res1, res2, res8, res8 and res4 in sequence, wherein DBL comprises convolutional layers, batch standardization and activation functions LeakyReLU, res1 comprises zero filling layers, DBL layers and 1 res unit layer, one res unit comprises two DBLs, res2 comprises a zero filling layer, a DBL layer and 2 res unit layers, res8 comprises a zero filling layer, a DBL layer and 8 res unit layers, and res4 comprises a zero filling layer, a DBL layer and 4 res unit layers;
res4 in Darknet-53 is output and connected with 5 DBLs, and after output, y1 is output after 1 DBL and 1 convolutional layer are obtained;
simultaneously, tensor splicing is carried out on the output through 1 DBL and upsampling and the output of the second res8 in Darknet-53, and y2 is output after 5 DBLs, 1 DBLs and convolutional layers;
meanwhile, the value after splicing and 5 DBLs is subjected to tensor splicing with the output of the 1 st res8 in Darknet-53 through 1 DBL and upsampling, and then the value after splicing is subjected to tensor splicing at the output y3 of 5 DBLs, 1 DBL and convolutional layer.
As a further scheme of the invention: the fusion ranging module determines whether the detection result is a violation building or not according to the context relationship, the distance relationship and the target size in the target detection result, so that the accuracy of target detection is improved.
As a further scheme of the invention: aerial photos of all road sections of roads near a scenic spot can be obtained in an aerial photographing mode, road picture samples of the roads near the scenic spot are obtained, pictures of suspected buildings against regulations are manually screened from the road picture samples, and the samples of the pictures against regulations are formed and used for training of a deep neural network model.
Compared with the prior art, the invention has the beneficial effects that: the method can automatically identify the buildings against the traffic regulations, improve the identification accuracy of the buildings against the traffic regulations, reduce the participation of manpower, and accelerate the identification efficiency so as to facilitate the on-site confirmation and supervision of the patrol personnel.
The method and the device have the advantages of realizing rapid and accurate identification of buildings violating the regulations, improving the identification efficiency of buildings violating the regulations in scenic spots, being well suitable for the change of external environments such as weather and the like, having better stability, having high operation efficiency, accurate detection, labor saving and strong robustness.
Drawings
Fig. 1 is a flow chart of the identification method of illegal buildings near scenic spots of the invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
As shown in fig. 1, a method for identifying illegal buildings near a scenic spot comprises the following steps:
constructing a deep neural network model, and training the constructed deep neural network model until the deep neural network model is trained to preset target parameters;
outputting the detected position area of each suspected violation building and the road segmentation mask of the road section to be detected of the target road by the trained deep neural network model; calculating the area of each suspected violation building and the distance between each suspected violation building and the road section to be detected of the target road; and presetting a distance discrimination threshold, and if the distance calculation value between the suspected violation building and the road section to be detected of the target road is smaller than the preset distance discrimination threshold for each suspected violation building, judging the suspected violation building as the violation building.
The method comprises the steps of firstly, acquiring data, acquiring aerial photos of all road sections of roads near a scenic spot in an aerial photographing mode, obtaining road picture samples of the roads near the scenic spot, manually screening out suspected illegal buildings from the road picture samples, and forming the illegal picture samples to be used for training a deep neural network model.
As a preferred implementation mode, the constructed deep neural network model comprises three modules, namely a road segmentation module, a violation building detection module and a fusion ranging module.
The road segmentation module adopts an FCN full convolution network, the structure of which is similar to a CNN structure of a convolution neural network, and the main difference is that a full connection layer is cancelled and replaced by a convolution layer mode, and finally a heat map is output. The FCN full convolutional network converts the last three fully-connected layers into three convolutional layers in the traditional convolutional neural network CNN structure. The first 5 layers are convolutional layers, the 6 th layer and the 7 th layer are respectively one-dimensional vectors with the length of 4096, the 8 th layer is one-dimensional vectors with the length of 1000 and respectively correspond to the probabilities of 1000 different categories, the FCN full convolutional network converts the 3 layers into convolutional layers, and the sizes (number of channels, width and height) of convolutional cores are respectively (4096, 1, 1), (1000, 1, 1).
The method for processing the original image by the road segmentation module comprises the following steps:
carrying out first vector convolution operation on the original image, and reducing the original image into 1/2 after pooling;
carrying out second vector convolution operation on the image, and reducing the image into 1/4 after pooling;
performing vector convolution operation and pooling for the third time on the image, reducing the image into 1/8 parts of the original image, and keeping the pooled feature map;
performing fourth vector convolution operation and pooling on the image, reducing the image to 1/16 of the original image, and keeping a pooled feature map;
and performing fifth vector convolution operation and pooling on the image, reducing the image into 1/32 of the original image, changing the full connection in the original CNN convolutional neural network operation into sixth vector convolution operation and seventh vector convolution operation, changing the number of feature maps of the image into 1/32 of the original image, obtaining an image heat map with a 1/32-size heat map, a 1/16-size feature map and a 1/8-size feature map, performing an up-sampling process on the 1/32-size heat map, and performing deconvolution on the image from the fifth vector convolution operation to the first vector convolution operation to supplement details.
The method for processing the image by the illegal building detection module comprises the following steps:
using yolov3 object detection technical feature extractor Darknet-53, wherein the main structures of 53 convolutional layers are DBL, res1, res2, res8, res8 and res4 in sequence, wherein DBL comprises convolutional layers, batch standardization and activation functions LeakyReLU, res1 comprises zero filling layers, DBL layers and 1 res unit layer, one res unit comprises two DBLs, res2 comprises a zero filling layer, a DBL layer and 2 res unit layers, res8 comprises a zero filling layer, a DBL layer and 8 res unit layers, and res4 comprises a zero filling layer, a DBL layer and 4 res unit layers;
res4 in Darknet-53 is output and connected with 5 DBLs, and after output, y1 is output after 1 DBL and 1 convolutional layer are obtained;
simultaneously, tensor splicing is carried out on the output through 1 DBL and upsampling and the output of the second res8 in Darknet-53, and y2 is output after 5 DBLs, 1 DBLs and convolutional layers;
meanwhile, the value after splicing and 5 DBLs is subjected to tensor splicing with the output of the 1 st res8 in Darknet-53 through 1 DBL and upsampling, and then the value after splicing is subjected to tensor splicing at the output y3 of 5 DBLs, 1 DBL and convolutional layer.
As a preferred implementation mode, the fusion ranging module determines whether the detection result is a violation building or not according to the context relationship, the distance relationship and the target size in the target detection result, so that the accuracy of target detection is improved.
The method can automatically identify the buildings against the traffic regulations, improve the identification accuracy of the buildings against the traffic regulations, reduce the participation of manpower, and accelerate the identification efficiency so as to facilitate the on-site confirmation and supervision of the patrol personnel.
The method and the device have the advantages of realizing rapid and accurate identification of buildings violating the regulations, improving the identification efficiency of buildings violating the regulations in scenic spots, being well suitable for the change of external environments such as weather and the like, having better stability, having high operation efficiency, accurate detection, labor saving and strong robustness.
It will be evident to those skilled in the art that the invention is not limited to the details of the foregoing illustrative embodiments, and that the present invention may be embodied in other specific forms without departing from the spirit or essential attributes thereof. The present embodiments are therefore to be considered in all respects as illustrative and not restrictive, the scope of the invention being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein. Any reference sign in a claim should not be construed as limiting the claim concerned.
Furthermore, it should be understood that although the present description refers to embodiments, not every embodiment may contain only a single embodiment, and such description is for clarity only, and those skilled in the art should integrate the description, and the embodiments may be combined as appropriate to form other embodiments understood by those skilled in the art.
Claims (6)
1. A method for identifying illegal buildings near a scenic spot is characterized by comprising the following steps:
constructing a deep neural network model, and training the constructed deep neural network model until the deep neural network model is trained to preset target parameters;
outputting the detected position area of each suspected violation building and the road segmentation mask of the road section to be detected of the target road by the trained deep neural network model; calculating the area of each suspected violation building and the distance between each suspected violation building and the road section to be detected of the target road; and presetting a distance discrimination threshold, and if the distance calculation value between the suspected violation building and the road section to be detected of the target road is smaller than the preset distance discrimination threshold for each suspected violation building, judging the suspected violation building as the violation building.
2. The method of identifying illegal buildings near scenic spots according to claim 1,
the constructed deep neural network model comprises three modules, namely a road segmentation module, a violation building detection module and a fusion ranging module.
3. The method of identifying a violation building near a scenic spot as recited in claim 2,
the method for processing the original image by the road segmentation module comprises the following steps:
carrying out first vector convolution operation on the original image, and reducing the original image into 1/2 after pooling;
carrying out second vector convolution operation on the image, and reducing the image into 1/4 after pooling;
performing vector convolution operation and pooling for the third time on the image, reducing the image into 1/8 parts of the original image, and keeping the pooled feature map;
performing fourth vector convolution operation and pooling on the image, reducing the image to 1/16 of the original image, and keeping a pooled feature map;
and performing fifth vector convolution operation and pooling on the image, reducing the image into 1/32 of the original image, changing the full connection in the original CNN convolutional neural network operation into sixth vector convolution operation and seventh vector convolution operation, changing the number of feature maps of the image into 1/32 of the original image, obtaining an image heat map with a 1/32-size heat map, a 1/16-size feature map and a 1/8-size feature map, performing an up-sampling process on the 1/32-size heat map, and performing deconvolution on the image from the fifth vector convolution operation to the first vector convolution operation to supplement details.
4. The method of identifying a violation building near a scenic spot as recited in claim 2,
the method for processing the image by the illegal building detection module comprises the following steps:
using yolov3 object detection technical feature extractor Darknet-53, wherein the main structures of 53 convolutional layers are DBL, res1, res2, res8, res8 and res4 in sequence, wherein DBL comprises convolutional layers, batch standardization and activation functions LeakyReLU, res1 comprises zero filling layers, DBL layers and 1 res unit layer, one res unit comprises two DBLs, res2 comprises a zero filling layer, a DBL layer and 2 res unit layers, res8 comprises a zero filling layer, a DBL layer and 8 res unit layers, and res4 comprises a zero filling layer, a DBL layer and 4 res unit layers;
res4 in Darknet-53 is output and connected with 5 DBLs, and after output, y1 is output after 1 DBL and 1 convolutional layer are obtained;
simultaneously, tensor splicing is carried out on the output through 1 DBL and upsampling and the output of the second res8 in Darknet-53, and y2 is output after 5 DBLs, 1 DBLs and convolutional layers;
meanwhile, the value after splicing and 5 DBLs is subjected to tensor splicing with the output of the 1 st res8 in Darknet-53 through 1 DBL and upsampling, and then the value after splicing is subjected to tensor splicing at the output y3 of 5 DBLs, 1 DBL and convolutional layer.
5. The method of identifying a violation building near a scenic spot as recited in claim 2,
the fusion ranging module determines whether the detection result is a violation building or not according to the context relationship, the distance relationship and the target size in the target detection result, so that the accuracy of target detection is improved.
6. The method of identifying illegal buildings near scenic spots according to claim 1,
aerial photos of all road sections of roads near a scenic spot can be obtained in an aerial photographing mode, road picture samples of the roads near the scenic spot are obtained, pictures of suspected buildings against regulations are manually screened from the road picture samples, and the samples of the pictures against regulations are formed and used for training of a deep neural network model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111316369.XA CN114005038A (en) | 2021-11-08 | 2021-11-08 | Method for identifying illegal buildings near scenic spot |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111316369.XA CN114005038A (en) | 2021-11-08 | 2021-11-08 | Method for identifying illegal buildings near scenic spot |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114005038A true CN114005038A (en) | 2022-02-01 |
Family
ID=79928148
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111316369.XA Pending CN114005038A (en) | 2021-11-08 | 2021-11-08 | Method for identifying illegal buildings near scenic spot |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114005038A (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104134350A (en) * | 2014-08-11 | 2014-11-05 | 浙江力石科技股份有限公司 | Intelligent dome camera system for traffic violation behavior recognition |
| US20180300549A1 (en) * | 2017-04-12 | 2018-10-18 | Baidu Online Network Technology (Beijing) Co., Ltd. | Road detecting method and apparatus |
| CN111209894A (en) * | 2020-02-10 | 2020-05-29 | 上海翼枭航空科技有限公司 | Roadside illegal building identification method for road aerial image |
| CN112307873A (en) * | 2020-07-08 | 2021-02-02 | 湖北大学 | Automatic illegal building identification method based on full convolution neural network |
| AU2020103901A4 (en) * | 2020-12-04 | 2021-02-11 | Chongqing Normal University | Image Semantic Segmentation Method Based on Deep Full Convolutional Network and Conditional Random Field |
| CN112613437A (en) * | 2020-12-28 | 2021-04-06 | 国网浙江省电力有限公司电力科学研究院 | High-accuracy illegal building identification method |
| CN113505643A (en) * | 2021-06-07 | 2021-10-15 | 浙江大华技术股份有限公司 | Violation target detection method and related device |
-
2021
- 2021-11-08 CN CN202111316369.XA patent/CN114005038A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104134350A (en) * | 2014-08-11 | 2014-11-05 | 浙江力石科技股份有限公司 | Intelligent dome camera system for traffic violation behavior recognition |
| US20180300549A1 (en) * | 2017-04-12 | 2018-10-18 | Baidu Online Network Technology (Beijing) Co., Ltd. | Road detecting method and apparatus |
| CN111209894A (en) * | 2020-02-10 | 2020-05-29 | 上海翼枭航空科技有限公司 | Roadside illegal building identification method for road aerial image |
| CN112307873A (en) * | 2020-07-08 | 2021-02-02 | 湖北大学 | Automatic illegal building identification method based on full convolution neural network |
| AU2020103901A4 (en) * | 2020-12-04 | 2021-02-11 | Chongqing Normal University | Image Semantic Segmentation Method Based on Deep Full Convolutional Network and Conditional Random Field |
| CN112613437A (en) * | 2020-12-28 | 2021-04-06 | 国网浙江省电力有限公司电力科学研究院 | High-accuracy illegal building identification method |
| CN113505643A (en) * | 2021-06-07 | 2021-10-15 | 浙江大华技术股份有限公司 | Violation target detection method and related device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11455805B2 (en) | Method and apparatus for detecting parking space usage condition, electronic device, and storage medium | |
| CN111582029B (en) | Traffic sign identification method based on dense connection and attention mechanism | |
| CN112329881B (en) | License plate recognition model training method, license plate recognition method and device | |
| CN113887412B (en) | Detection method, detection terminal, monitoring system and storage medium for pollution emission | |
| CN111209894A (en) | Roadside illegal building identification method for road aerial image | |
| CN110852164A (en) | YOLOv 3-based method and system for automatically detecting illegal building | |
| CN110490150A (en) | A kind of automatic auditing system of picture violating the regulations and method based on vehicle retrieval | |
| CN110796580A (en) | Intelligent traffic system management method and related products | |
| CN110087041A (en) | Video data processing and transmission method and system based on the base station 5G | |
| CN114898243A (en) | Traffic scene analysis method and device based on video stream | |
| CN114926791A (en) | Method and device for detecting abnormal lane change of vehicles at intersection, storage medium and electronic equipment | |
| CN110826364A (en) | Stock position identification method and device | |
| CN112560814A (en) | Method for identifying vehicles entering and exiting parking spaces | |
| CN116229396B (en) | High-speed pavement disease identification and warning method | |
| CN112926415A (en) | Pedestrian avoiding system and pedestrian monitoring method | |
| CN114005038A (en) | Method for identifying illegal buildings near scenic spot | |
| CN110765900A (en) | DSSD-based automatic illegal building detection method and system | |
| CN116129327A (en) | Infrared vehicle detection method based on improved YOLOv7 algorithm | |
| CN115359392A (en) | Roadside construction behavior detection method based on improved YOLOv5 model | |
| CN113887420A (en) | AI-based intelligent detection and identification system for urban public parking spaces | |
| CN113011338A (en) | Lane line detection method and system | |
| CN114627400A (en) | Lane congestion detection method and device, electronic equipment and storage medium | |
| CN111340001A (en) | Traffic jam identification method based on remote sensing image deep learning | |
| CN112767469B (en) | Highly intelligent acquisition method for urban mass buildings | |
| CN114241792B (en) | Traffic flow detection method and system |
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 |