CN117475327A - Multi-target detection positioning method and system based on remote sensing image in city - Google Patents
Multi-target detection positioning method and system based on remote sensing image in city Download PDFInfo
- Publication number
- CN117475327A CN117475327A CN202311625120.6A CN202311625120A CN117475327A CN 117475327 A CN117475327 A CN 117475327A CN 202311625120 A CN202311625120 A CN 202311625120A CN 117475327 A CN117475327 A CN 117475327A
- Authority
- CN
- China
- Prior art keywords
- remote sensing
- image
- information
- clear
- sensing image
- 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.)
- Withdrawn
Links
- 238000000034 method Methods 0.000 title claims abstract description 88
- 238000001514 detection method Methods 0.000 title claims abstract description 37
- 238000012545 processing Methods 0.000 claims abstract description 46
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 22
- 238000000605 extraction Methods 0.000 claims abstract description 20
- 238000007781 pre-processing Methods 0.000 claims abstract description 19
- 230000008569 process Effects 0.000 claims description 23
- 238000005457 optimization Methods 0.000 claims description 12
- 238000007477 logistic regression Methods 0.000 claims description 10
- 238000012706 support-vector machine Methods 0.000 claims description 10
- 238000012549 training Methods 0.000 claims description 9
- 238000004422 calculation algorithm Methods 0.000 claims description 8
- 238000001914 filtration Methods 0.000 claims description 8
- 238000011176 pooling Methods 0.000 claims description 6
- 230000004913 activation Effects 0.000 claims description 3
- 238000010606 normalization Methods 0.000 claims description 3
- 238000001308 synthesis method Methods 0.000 claims description 3
- 230000015572 biosynthetic process Effects 0.000 description 4
- 230000006870 function Effects 0.000 description 4
- 238000003786 synthesis reaction Methods 0.000 description 4
- 238000004458 analytical method Methods 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- 230000003044 adaptive effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000007635 classification algorithm Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
- G06V20/13—Satellite images
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/0464—Convolutional networks [CNN, ConvNet]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/09—Supervised learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/22—Image preprocessing by selection of a specific region containing or referencing a pattern; Locating or processing of specific regions to guide the detection or recognition
- G06V10/225—Image preprocessing by selection of a specific region containing or referencing a pattern; Locating or processing of specific regions to guide the detection or recognition based on a marking or identifier characterising the area
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
- G06V10/443—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components by matching or filtering
- G06V10/449—Biologically inspired filters, e.g. difference of Gaussians [DoG] or Gabor filters
- G06V10/451—Biologically inspired filters, e.g. difference of Gaussians [DoG] or Gabor filters with interaction between the filter responses, e.g. cortical complex cells
- G06V10/454—Integrating the filters into a hierarchical structure, e.g. convolutional neural networks [CNN]
-
- 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
- G06V10/74—Image or video pattern matching; Proximity measures in feature spaces
- G06V10/761—Proximity, similarity or dissimilarity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/77—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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/10—Terrestrial scenes
- G06V20/17—Terrestrial scenes taken from planes or by drones
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/07—Target detection
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Health & Medical Sciences (AREA)
- Software Systems (AREA)
- Computing Systems (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Medical Informatics (AREA)
- Databases & Information Systems (AREA)
- Molecular Biology (AREA)
- Biomedical Technology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Data Mining & Analysis (AREA)
- Remote Sensing (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Astronomy & Astrophysics (AREA)
- Biodiversity & Conservation Biology (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a multi-target detection positioning method based on remote sensing images in cities, which comprises the following steps of: step one: acquiring remote sensing image information in a city, and preprocessing the remote sensing image to acquire a clear remote sensing image; step two: feature extraction, namely performing feature extraction on a clear remote sensing image by using a convolutional neural network, and extracting effective feature information from the image; step three: the method comprises the steps of detecting targets, classifying the extracted features by using a classifier, identifying a plurality of targets in an image, positioning the targets and obtaining target positioning information; step four: after the target positioning information is acquired, acquiring interval information among the target positioning information, and processing the interval information among the target positioning information to select an auxiliary positioning position; the invention can obtain more accurate and reliable multi-target detection and positioning, and meets different use requirements of users.
Description
Technical Field
The invention relates to the technical field of positioning systems, in particular to a multi-target detection positioning method and system based on remote sensing images in cities.
Background
The multi-target detection positioning of the remote sensing image refers to detecting and positioning a plurality of targets from the remote sensing image, and in the multi-target detection positioning of the remote sensing image, the following points need to be noted: data quality, feature selection, model selection, result evaluation and the like;
the multi-target detection positioning method and system can be used in multi-target detection positioning of remote sensing images.
In the existing multi-target detection positioning method and system, the type of the acquired data is single in the positioning process, so that the accuracy and reliability of the finally analyzed detection system information are not high enough, and a certain image is brought to the use of the multi-target detection positioning method and system, so that the multi-target detection positioning method and system based on remote sensing images in cities are provided.
Disclosure of Invention
Aiming at the defects in the prior art, the invention provides a multi-target detection positioning method based on remote sensing images in cities, which comprises the following steps:
step one: acquiring remote sensing image information in a city, and preprocessing the remote sensing image to acquire a clear remote sensing image;
step two: feature extraction, namely performing feature extraction on a clear remote sensing image by using a convolutional neural network, and extracting effective feature information from the image;
step three: the method comprises the steps of detecting targets, classifying the extracted features by using a classifier, identifying a plurality of targets in an image, positioning the targets and obtaining target positioning information;
step four: after the target positioning information is acquired, acquiring interval information among the target positioning information, and processing the interval information among the target positioning information to select an auxiliary positioning position;
step five: after the auxiliary positioning position is selected, the auxiliary positioning position and each piece of target positioning information are processed to obtain auxiliary positioning information;
step six, a step of performing a step of; and simultaneously deriving final positioning information from the target positioning information and the auxiliary positioning information, and then transmitting the final positioning information to a preset receiving terminal.
The method for acquiring the remote sensing image information in the first step comprises aerial shooting, satellite remote sensing and unmanned aerial vehicle remote sensing, wherein the image information acquired by aerial shooting is marked as a first remote sensing image, the image acquired by satellite remote sensing is marked as a second remote sensing image, and the image information acquired by unmanned aerial vehicle remote sensing is marked as a third remote sensing image;
preprocessing the first remote sensing image, the second remote sensing image and the third remote sensing image, namely performing definition optimization processing on the first remote sensing image, the second remote sensing image and the third remote sensing image to obtain a first clear image, a second clear image and a third clear image, and processing the first clear image, the second clear image and the third clear image to select a final image, namely selecting the clear remote sensing image.
The specific process of performing definition optimization processing on the first remote sensing image, the second remote sensing image and the third remote sensing image is as follows: the method comprises the steps of importing a first remote sensing image into a processing library, performing image enhancement by setting an image interpolation method, a sharpening filtering method, a super-resolution technology, a denoising method, a contrast enhancement method and an HDR synthesis method in the processing library, randomly selecting three methods from the processing library, processing the first remote sensing image to obtain three processed first remote sensing images, performing definition comparison on the three processed first remote sensing images, and selecting the processed first remote sensing image with highest definition as a first clear image;
the process of performing definition optimization processing on the second remote sensing image and the third remote sensing image is the same as the process of processing the first remote sensing image;
the first clear image, the second clear image and the third clear image are processed to select a final image, namely, the specific process of selecting the clear remote sensing image is as follows: comparing the similarity of the first clear image, the second clear image and the third clear image, and extracting the highest definition of the first clear image, the second clear image and the third clear image as a selected clear remote sensing image when the similarity of the first clear image, the second clear image and the third clear image is larger than a preset value;
and when the similarity of the first clear image, the second clear image and the third clear image is smaller than the preset value, the first clear image, the second clear image and the third clear image are selected again.
The specific process of extracting the characteristics of the clear remote sensing image by using the convolutional neural network is as follows:
s1: data preprocessing: firstly, converting a clear remote sensing image into a preset format, wherein the preset format comprises JPEG and PNG, secondly, carrying out normalization processing on the image, converting pixel values into a range of 0-1, and then carrying out data enhancement operation, wherein the data enhancement operation comprises random cutting, rotation and overturning, so as to obtain the preprocessed clear remote sensing image;
s2: constructing a convolutional neural network model: the convolutional neural network consists of a plurality of convolutional layers, a pooling layer and a full-connection layer, wherein the convolutional layers extract local features of images through convolutional operation, the pooling layer reduces the sizes of feature images through downsampling operation, the full-connection layer classifies through connecting all the feature images, when a model is built, the number of layers of the network, the size of a convolutional kernel of each layer and activation function parameters are determined, and a loss function and an optimization algorithm are selected according to task requirements;
s3: training a model: and inputting the preprocessed remote sensing image data into the constructed convolutional neural network model, and training the model. In the training process, the network automatically learns the characteristic representation according to the input image data;
s4: feature extraction: the trained convolutional neural network performs feature extraction on the input preprocessed clear remote sensing image;
s5: outputting a result: and extracting required identification characteristics, and identifying a corresponding target.
The classifier in the third step comprises a support vector machine, a logistic regression and a softmax classifier, and one of the support vector machine, the logistic regression and the softmax classifier is selected randomly each time the classifier is used for classifying the extracted features.
The specific processing procedure for processing the interval information among the target positioning information to select the auxiliary positioning position is as follows: extracting interval information among the target positioning information, marking the number of the target positioning information as i, marking the interval information among the target positioning information as K, extracting two target positioning information corresponding to the maximum value in the interval information K among the i target positioning information, connecting the two points to obtain an auxiliary line L, then selecting the highest building and the second high building closest to the auxiliary line L from the clear remote sensing image, marking the highest building and the second high building closest to the auxiliary line L as an auxiliary point A and an auxiliary point B, and forming an auxiliary positioning position by the auxiliary point A and the auxiliary point B.
The specific processing procedure for processing the auxiliary positioning position and each piece of target positioning information to obtain the auxiliary positioning information is as follows: and extracting the obtained auxiliary positioning position, and measuring the distance information between the auxiliary positioning position and each piece of target positioning information, namely obtaining the auxiliary positioning information.
The multi-target detection positioning system based on the remote sensing image in the city comprises a remote sensing image acquisition module, a preprocessing module, a data characteristic extraction module, a target detection module, an auxiliary positioning module and an information sending module;
the remote sensing image acquisition module is used for acquiring remote sensing image information, wherein the remote sensing image information comprises an aerial shooting image, a satellite remote sensing image and an unmanned aerial vehicle remote sensing image;
the preprocessing module is used for preprocessing the remote sensing image to obtain a clear remote sensing image;
the data characteristic extraction module is used for extracting characteristics of the clear remote sensing image by utilizing the convolutional neural network, and extracting effective characteristic information from the image;
the target detection module is used for classifying the extracted features by using a classifier, identifying a plurality of targets in the image, positioning the targets and obtaining target positioning information;
the auxiliary positioning module is used for performing auxiliary positioning to acquire auxiliary positioning information;
the information sending module is used for simultaneously exporting final positioning information from the target positioning information and the auxiliary positioning information, and then sending the final positioning information to a preset receiving terminal.
The beneficial effects of the invention are as follows:
according to the method, different types of urban remote sensing image information are collected in multiple modes at the same time, so that better data support is provided for subsequent multi-target detection positioning, more types of remote sensing image data are optimized, clearer remote sensing image information can be obtained, and finally the obtained positioning information is more accurate, so that the positioning information positioned by the method is smaller in error and higher in reliability, meanwhile, the urban remote sensing image positioning also has the advantages of large monitoring range, real-time transmission, less limitation on ground conditions, large information quantity, multiple means and the like, auxiliary positioning information is further provided for a user to provide positioning reference, the user can better know the position information of multiple targets, and comprehensive multi-target detection positioning is realized.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below. Like elements or portions are generally identified by like reference numerals throughout the several figures. In the drawings, elements or portions thereof are not necessarily drawn to scale.
FIG. 1 is an overall flow chart of the method of the present invention;
fig. 2 is a block diagram of the system of the present invention.
Detailed Description
Embodiments of the technical scheme of the present invention will be described in detail below with reference to the accompanying drawings. The following examples are only for more clearly illustrating the technical aspects of the present invention, and thus are merely examples, and are not intended to limit the scope of the present invention.
It is noted that unless otherwise indicated, technical or scientific terms used herein should be given the ordinary meaning as understood by one of ordinary skill in the art to which this invention pertains.
As shown in fig. 1, a multi-target detection positioning method based on remote sensing images in a city comprises the following steps:
step one: acquiring remote sensing image information in a city, and preprocessing the remote sensing image to acquire a clear remote sensing image;
step two: feature extraction, namely performing feature extraction on a clear remote sensing image by using a convolutional neural network, and extracting effective feature information from the image;
step three: the method comprises the steps of detecting targets, classifying the extracted features by using a classifier, identifying a plurality of targets in an image, positioning the targets and obtaining target positioning information;
step four: after the target positioning information is acquired, acquiring interval information among the target positioning information, and processing the interval information among the target positioning information to select an auxiliary positioning position;
step five: after the auxiliary positioning position is selected, the auxiliary positioning position and each piece of target positioning information are processed to obtain auxiliary positioning information;
step six, a step of performing a step of; and simultaneously deriving final positioning information from the target positioning information and the auxiliary positioning information, and then transmitting the final positioning information to a preset receiving terminal.
According to the method, different types of urban remote sensing image information are collected in multiple modes at the same time, so that better data support is provided for subsequent multi-target detection positioning, more types of remote sensing image data are optimized, clearer remote sensing image information can be obtained, and finally the obtained positioning information is more accurate, so that the positioning information positioned by the method is smaller in error and higher in reliability, meanwhile, the urban remote sensing image positioning also has the advantages of large monitoring range, real-time transmission, less limitation on ground conditions, large information quantity, multiple means and the like, auxiliary positioning information is further provided for a user to provide positioning reference, the user can better know the position information of multiple targets, and comprehensive multi-target detection positioning is realized.
The method for acquiring the remote sensing image information in the first step comprises aerial shooting, satellite remote sensing and unmanned aerial vehicle remote sensing, wherein the image information acquired by aerial shooting is marked as a first remote sensing image, the image acquired by satellite remote sensing is marked as a second remote sensing image, and the image information acquired by unmanned aerial vehicle remote sensing is marked as a third remote sensing image;
preprocessing the first remote sensing image, the second remote sensing image and the third remote sensing image, namely performing definition optimization processing on the first remote sensing image, the second remote sensing image and the third remote sensing image to obtain a first clear image, a second clear image and a third clear image, and processing the first clear image, the second clear image and the third clear image to select a final image, namely selecting a clear remote sensing image;
through the process, the remote sensing image information acquired in different modes is intelligently processed, the definition of the acquired remote sensing images in different selected modes is ensured, and data support is provided for subsequent positioning analysis.
The specific process of performing definition optimization processing on the first remote sensing image, the second remote sensing image and the third remote sensing image is as follows: the method comprises the steps of importing a first remote sensing image into a processing library, performing image enhancement by setting an image interpolation method, a sharpening filtering method, a super-resolution technology, a denoising method, a contrast enhancement method and an HDR synthesis method in the processing library, randomly selecting three methods from the processing library, processing the first remote sensing image to obtain three processed first remote sensing images, performing definition comparison on the three processed first remote sensing images, and selecting the processed first remote sensing image with highest definition as a first clear image;
the process of performing definition optimization processing on the second remote sensing image and the third remote sensing image is the same as the process of processing the first remote sensing image;
the first clear image, the second clear image and the third clear image are processed to select a final image, namely, the specific process of selecting the clear remote sensing image is as follows: comparing the similarity of the first clear image, the second clear image and the third clear image, and extracting the highest definition of the first clear image, the second clear image and the third clear image as a selected clear remote sensing image when the similarity of the first clear image, the second clear image and the third clear image is larger than a preset value;
when the similarity of the first clear image, the second clear image and the third clear image is smaller than a preset value, the first clear image, the second clear image and the third clear image are selected again;
through the process, the remote sensing image with the highest definition is selected as the final positioning analysis image information, so that the accuracy of the analyzed positioning information is ensured;
image interpolation: this is a technique of reassigning image pixels by a mathematical algorithm to increase resolution and sharpness of an image. Typical interpolation algorithms include nearest neighbor interpolation, bilinear interpolation, bicubic interpolation, and the like.
Sharpening filtering: sharpening filtering makes the image appear clearer by enhancing the edges and contours of the image. Common sharpening filters include laplace sharpening, high pass filtering, gradient sharpening, and the like.
Super resolution technology: the super-resolution technology learns and deduces the detail and texture information of the high-resolution image by analyzing the similarity between the low-resolution image and the corresponding high-resolution image, thereby improving the definition of the image. Common super-resolution techniques include super-resolution techniques based on deep learning, super-resolution techniques based on interpolation, and the like.
Denoising: for some images containing noise, the influence of the noise on the images can be reduced by a denoising technology, so that the definition of the images is improved. Common denoising algorithms include median filtering, gaussian filtering, wavelet denoising, and the like.
Contrast enhancement: by increasing the contrast of the image, the details in the image can be more prominent, thereby improving the definition of the image. Common contrast enhancement algorithms are histogram equalization, adaptive histogram equalization, contrast stretching, etc.
HDR synthesis: HDR (high dynamic range) synthesis is the synthesis of multiple images at different exposure levels, resulting in an image that contains more detail and dynamic range. The definition and visual effect of the image can be improved through HDR synthesis;
these methods may be selected for use according to specific requirements and conditions, but it should be noted that some new distortions or noise may be introduced during the process of performing sharpness improvement, and the selection needs to be made according to practical situations.
The specific process of extracting the characteristics of the clear remote sensing image by using the convolutional neural network is as follows:
s1: data preprocessing: firstly, converting a clear remote sensing image into a preset format, wherein the preset format comprises JPEG and PNG, secondly, carrying out normalization processing on the image, converting pixel values into a range of 0-1, and then carrying out data enhancement operation, wherein the data enhancement operation comprises random cutting, rotation and overturning, so as to obtain the preprocessed clear remote sensing image;
s2: constructing a convolutional neural network model: the convolutional neural network consists of a plurality of convolutional layers, a pooling layer and a full-connection layer, wherein the convolutional layers extract local features of images through convolutional operation, the pooling layer reduces the sizes of feature images through downsampling operation, the full-connection layer classifies through connecting all the feature images, when a model is built, the number of layers of the network, the size of a convolutional kernel of each layer and activation function parameters are determined, and a loss function and an optimization algorithm are selected according to task requirements;
s3: training a model: and inputting the preprocessed remote sensing image data into the constructed convolutional neural network model, and training the model. In the training process, the network automatically learns the characteristic representation according to the input image data;
s4: feature extraction: the trained convolutional neural network performs feature extraction on the input preprocessed clear remote sensing image;
s5: outputting a result: and extracting required identification characteristics, and identifying a corresponding target.
The classifier in the third step comprises a support vector machine, a logistic regression and a softmax classifier, and when the classifier is used for classifying the extracted features each time, one of the support vector machine, the logistic regression and the softmax classifier is selected randomly;
support Vector Machine (SVM): SVM is a supervised learning algorithm that can be used for classification and regression problems. In object detection, SVM may be used to classify different object objects.
Logistic regression (Logistic Regression): logistic regression is a probabilistic model that can be used to classify problems. In target detection, logistic regression may be used to determine the class of target objects.
softmax classifier: the softmax classifier is a multi-classification algorithm that may be used to classify a target object. The softmax classifier maps the input features to a plurality of categories, assigns a probability value to each category, and finally outputs the probability that the target object belongs to each category.
The specific processing procedure for processing the interval information among the target positioning information to select the auxiliary positioning position is as follows: extracting interval information among the target positioning information, marking the number of the target positioning information as i, marking the interval information among the target positioning information as K, extracting two target positioning information corresponding to the maximum value in the interval information K among the i target positioning information, connecting the two points to obtain an auxiliary line L, then selecting the highest building and the second high building closest to the auxiliary line L from the clear remote sensing image, marking the highest building and the second high building closest to the auxiliary line L as an auxiliary point A and an auxiliary point B, and forming an auxiliary positioning position by the auxiliary point A and the auxiliary point B.
The specific processing procedure for processing the auxiliary positioning position and each target positioning information to obtain the auxiliary positioning information is as follows: extracting and acquiring auxiliary positioning positions, and measuring distance information between the auxiliary positioning positions and each piece of target positioning information, namely acquiring the auxiliary positioning information;
through the process, more accurate auxiliary positioning information is obtained, so that a user can more accurately know the positioning of a plurality of targets.
As shown in fig. 2, a multi-target detection positioning system based on remote sensing images in cities comprises a remote sensing image acquisition module, a preprocessing module, a data feature extraction module, a target detection module, an auxiliary positioning module and an information sending module;
the remote sensing image acquisition module is used for acquiring remote sensing image information, wherein the remote sensing image information comprises an aerial shooting image, a satellite remote sensing image and an unmanned aerial vehicle remote sensing image;
the preprocessing module is used for preprocessing the remote sensing image to obtain a clear remote sensing image;
the data characteristic extraction module is used for extracting characteristics of the clear remote sensing image by utilizing the convolutional neural network, and extracting effective characteristic information from the image;
the target detection module is used for classifying the extracted features by using a classifier, identifying a plurality of targets in the image, positioning the targets and obtaining target positioning information;
the auxiliary positioning module is used for performing auxiliary positioning to acquire auxiliary positioning information;
the information sending module is used for simultaneously exporting final positioning information from the target positioning information and the auxiliary positioning information, and then sending the final positioning information to a preset receiving terminal.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present invention, and not for limiting the same; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some or all of the technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit of the invention, and are intended to be included within the scope of the appended claims and description.
Claims (8)
1. A multi-target detection positioning method based on remote sensing images in cities is characterized by comprising the following steps:
step one: acquiring remote sensing image information in a city, and preprocessing the remote sensing image to acquire a clear remote sensing image;
step two: feature extraction, namely performing feature extraction on a clear remote sensing image by using a convolutional neural network, and extracting effective feature information from the image;
step three: the method comprises the steps of detecting targets, classifying the extracted features by using a classifier, identifying a plurality of targets in an image, positioning the targets and obtaining target positioning information;
step four: after the target positioning information is acquired, acquiring interval information among the target positioning information, and processing the interval information among the target positioning information to select an auxiliary positioning position;
step five: after the auxiliary positioning position is selected, the auxiliary positioning position and each piece of target positioning information are processed to obtain auxiliary positioning information;
step six, a step of performing a step of; and simultaneously deriving final positioning information from the target positioning information and the auxiliary positioning information, and then transmitting the final positioning information to a preset receiving terminal.
2. The method for multi-target detection and positioning based on remote sensing images in cities according to claim 1, wherein the method comprises the following steps: the method for acquiring the remote sensing image information in the first step comprises aerial shooting, satellite remote sensing and unmanned aerial vehicle remote sensing, wherein the image information acquired by aerial shooting is marked as a first remote sensing image, the image acquired by satellite remote sensing is marked as a second remote sensing image, and the image information acquired by unmanned aerial vehicle remote sensing is marked as a third remote sensing image;
preprocessing the first remote sensing image, the second remote sensing image and the third remote sensing image, namely performing definition optimization processing on the first remote sensing image, the second remote sensing image and the third remote sensing image to obtain a first clear image, a second clear image and a third clear image, and processing the first clear image, the second clear image and the third clear image to select a final image, namely selecting the clear remote sensing image.
3. The method for multi-target detection and positioning based on remote sensing images in cities according to claim 2, wherein the method comprises the following steps: the specific process of performing definition optimization processing on the first remote sensing image, the second remote sensing image and the third remote sensing image is as follows: the method comprises the steps of importing a first remote sensing image into a processing library, performing image enhancement by setting an image interpolation method, a sharpening filtering method, a super-resolution technology, a denoising method, a contrast enhancement method and an HDR synthesis method in the processing library, randomly selecting three methods from the processing library, processing the first remote sensing image to obtain three processed first remote sensing images, performing definition comparison on the three processed first remote sensing images, and selecting the processed first remote sensing image with highest definition as a first clear image;
the process of performing definition optimization processing on the second remote sensing image and the third remote sensing image is the same as the process of processing the first remote sensing image;
the first clear image, the second clear image and the third clear image are processed to select a final image, namely, the specific process of selecting the clear remote sensing image is as follows: comparing the similarity of the first clear image, the second clear image and the third clear image, and extracting the highest definition of the first clear image, the second clear image and the third clear image as a selected clear remote sensing image when the similarity of the first clear image, the second clear image and the third clear image is larger than a preset value;
and when the similarity of the first clear image, the second clear image and the third clear image is smaller than the preset value, the first clear image, the second clear image and the third clear image are selected again.
4. The method for multi-target detection and positioning based on remote sensing images in cities according to claim 1, wherein the method comprises the following steps: the specific process of extracting the characteristics of the clear remote sensing image by using the convolutional neural network is as follows:
s1: data preprocessing: firstly, converting a clear remote sensing image into a preset format, wherein the preset format comprises JPEG and PNG, secondly, carrying out normalization processing on the image, converting pixel values into a range of 0-1, and then carrying out data enhancement operation, wherein the data enhancement operation comprises random cutting, rotation and overturning, so as to obtain the preprocessed clear remote sensing image;
s2: constructing a convolutional neural network model: the convolutional neural network consists of a plurality of convolutional layers, a pooling layer and a full-connection layer, wherein the convolutional layers extract local features of images through convolutional operation, the pooling layer reduces the sizes of feature images through downsampling operation, the full-connection layer classifies through connecting all the feature images, when a model is built, the number of layers of the network, the size of a convolutional kernel of each layer and activation function parameters are determined, and a loss function and an optimization algorithm are selected according to task requirements;
s3: training a model: inputting the preprocessed remote sensing image data into a constructed convolutional neural network model, and training the model, wherein in the training process, the network automatically learns characteristic representation according to the input image data;
s4: feature extraction: the trained convolutional neural network performs feature extraction on the input preprocessed clear remote sensing image;
s5: outputting a result: and extracting required identification characteristics, and identifying a corresponding target.
5. The method for multi-target detection and positioning based on remote sensing images in cities according to claim 1, wherein the method comprises the following steps: the classifier in the third step comprises a support vector machine, a logistic regression and a softmax classifier, and one of the support vector machine, the logistic regression and the softmax classifier is randomly selected each time the classifier is used for classifying the extracted features.
6. The method for multi-target detection and positioning based on remote sensing images in cities according to claim 1, wherein the method comprises the following steps: the specific processing procedure for processing the interval information among the target positioning information to select the auxiliary positioning position is as follows: extracting interval information among the target positioning information, marking the number of the target positioning information as i, marking the interval information among the target positioning information as K, extracting two target positioning information corresponding to the maximum value in the interval information K among the i target positioning information, connecting the two points to obtain an auxiliary line L, then selecting the highest building and the second high building closest to the auxiliary line L from the clear remote sensing image, marking the highest building and the second high building closest to the auxiliary line L as an auxiliary point A and an auxiliary point B, and forming an auxiliary positioning position by the auxiliary point A and the auxiliary point B.
7. The method for multi-target detection and positioning based on remote sensing images in cities according to claim 1, wherein the method comprises the following steps: the specific processing procedure for processing the auxiliary positioning position and each target positioning information to obtain the auxiliary positioning information is as follows: and extracting the obtained auxiliary positioning position, and measuring the distance information between the auxiliary positioning position and each piece of target positioning information, namely obtaining the auxiliary positioning information.
8. A multi-target detection positioning system based on remote sensing images in a city, wherein the positioning system is applied to the positioning method of any one of claims 1 to 7, and is characterized in that: the positioning system comprises a remote sensing image acquisition module, a preprocessing module, a data characteristic extraction module, a target detection module, an auxiliary positioning module and an information sending module;
the remote sensing image acquisition module is used for acquiring remote sensing image information, wherein the remote sensing image information comprises an aerial shooting image, a satellite remote sensing image and an unmanned aerial vehicle remote sensing image;
the preprocessing module is used for preprocessing the remote sensing image to obtain a clear remote sensing image;
the data characteristic extraction module is used for extracting characteristics of the clear remote sensing image by utilizing the convolutional neural network, and extracting effective characteristic information from the image;
the target detection module is used for classifying the extracted features by using a classifier, identifying a plurality of targets in the image, positioning the targets and obtaining target positioning information;
the auxiliary positioning module is used for performing auxiliary positioning to acquire auxiliary positioning information;
the information sending module is used for simultaneously exporting final positioning information from the target positioning information and the auxiliary positioning information, and then sending the final positioning information to a preset receiving terminal.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311625120.6A CN117475327A (en) | 2023-11-30 | 2023-11-30 | Multi-target detection positioning method and system based on remote sensing image in city |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311625120.6A CN117475327A (en) | 2023-11-30 | 2023-11-30 | Multi-target detection positioning method and system based on remote sensing image in city |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117475327A true CN117475327A (en) | 2024-01-30 |
Family
ID=89632957
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311625120.6A Withdrawn CN117475327A (en) | 2023-11-30 | 2023-11-30 | Multi-target detection positioning method and system based on remote sensing image in city |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117475327A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118071877A (en) * | 2024-04-19 | 2024-05-24 | 武汉追月信息技术有限公司 | Urban mapping service method and system based on remote sensing image |
-
2023
- 2023-11-30 CN CN202311625120.6A patent/CN117475327A/en not_active Withdrawn
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118071877A (en) * | 2024-04-19 | 2024-05-24 | 武汉追月信息技术有限公司 | Urban mapping service method and system based on remote sensing image |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112418117B (en) | Small target detection method based on unmanned aerial vehicle image | |
| CN111310666B (en) | High-resolution image ground feature identification and segmentation method based on texture features | |
| CN111079640B (en) | Vehicle type identification method and system based on automatic amplification sample | |
| CN109766858A (en) | Three-dimensional convolution neural network hyperspectral image classification method combined with bilateral filtering | |
| CN105809121A (en) | Multi-characteristic synergic traffic sign detection and identification method | |
| CN112200121B (en) | Hyperspectral unknown target detection method based on EVM and deep learning | |
| CN108765465A (en) | A kind of unsupervised SAR image change detection | |
| CN115439458A (en) | Industrial image defect target detection algorithm based on depth map attention | |
| CN116012291A (en) | Industrial part image defect detection method and system, electronic equipment and storage medium | |
| CN111915583A (en) | Vehicle and pedestrian detection method based on vehicle-mounted thermal infrared imager in complex scene | |
| CN115063786A (en) | High-order distant view fuzzy license plate detection method | |
| CN110070545B (en) | Method for automatically extracting urban built-up area by urban texture feature density | |
| CN113421223B (en) | Industrial product surface defect detection method based on deep learning and Gaussian mixture | |
| CN111814852A (en) | Image detection method, image detection device, electronic equipment and computer-readable storage medium | |
| CN111639697B (en) | Hyperspectral image classification method based on non-repeated sampling and prototype network | |
| CN117475327A (en) | Multi-target detection positioning method and system based on remote sensing image in city | |
| CN115409789A (en) | Power transmission line engineering defect detection method based on image semantic segmentation | |
| CN117315670B (en) | Water meter reading area detection method based on computer vision | |
| CN111507258B (en) | Road area detection method and system based on ensemble learning for focusing key samples | |
| Roy | Automatics number plate recognition using convolution neural network | |
| CN114648738A (en) | Image identification system and method based on Internet of things and edge calculation | |
| CN115393743A (en) | Vehicle detection method based on double-branch encoding and decoding network, unmanned aerial vehicle and medium | |
| CN114862883A (en) | Target edge extraction method, image segmentation method and system | |
| CN114140698A (en) | Water system information extraction algorithm based on FasterR-CNN | |
| CN118351118B (en) | Printer shell defect detection method and device based on deep convolutional neural network |
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 | ||
| WW01 | Invention patent application withdrawn after publication | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20240130 |