CN115965639A - Intelligent water conservancy image processing method, device and system - Google Patents
Intelligent water conservancy image processing method, device and system Download PDFInfo
- Publication number
- CN115965639A CN115965639A CN202211676362.3A CN202211676362A CN115965639A CN 115965639 A CN115965639 A CN 115965639A CN 202211676362 A CN202211676362 A CN 202211676362A CN 115965639 A CN115965639 A CN 115965639A
- Authority
- CN
- China
- Prior art keywords
- image
- area
- ratio
- detected
- predicted
- 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
Images
Classifications
-
- 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
- Y02A90/30—Assessment of water resources
Abstract
The invention provides an intelligent water conservancy image processing method, device and system, wherein the method comprises the following steps: acquiring a target image, preprocessing the target image, dividing image sub-blocks, and determining the area where the image sub-blocks are located; acquiring an image set in a database; carrying out binarization processing on images in the image set to obtain a binarization image set, and determining a first land contour and a first water area contour; acquiring an image to be detected, wherein the image to be detected and the image set have the same shooting area; carrying out binarization processing on an image to be detected to obtain a binarized image to be detected; determining a second land profile and a second water area profile according to the binaryzation image to be detected; calculating a first ratio and a second ratio; and when the first ratio and the second ratio meet the preset condition, determining the predicted water level, and when the predicted water level exceeds a preset water level threshold, starting an emergency plan for effectively detecting and identifying the abnormal condition of the water level.
Description
Technical Field
The invention relates to the technical field of image processing, in particular to an intelligent water conservancy image processing method, device and system.
Background
With the development of science and technology, monitoring systems for preventing flood disasters gradually tend to be digital and intelligent, effective monitoring of all water areas is difficult to achieve for areas with numerous water systems and wide regions, when a flood season comes, loss can be reduced only through a passive disaster prevention method, and an effective prediction method for natural disasters is lacked. In which historical data may lose reference value over time, leading to increased difficulty in detection and identification of natural disasters.
In the prior art, digitization and informatization of the water conservancy industry are in an exploration stage, technical means are single, monitoring is carried out through the internet of things formed by a simple sensing system and network equipment with low power consumption, the monitoring means is single, the monitoring effect is poor, manpower is consumed for active monitoring, and the existing image identification method is poor in adaptability to characteristics changing in the field, so that abnormal information cannot be accurately captured when abnormality occurs, and the identification effect is poor.
Disclosure of Invention
The invention solves the problem of how to effectively monitor and identify the abnormal condition of the water level.
In order to solve the above problems, the present invention provides a method, an apparatus and a system for processing an intelligent water conservancy image.
In a first aspect, the present invention provides an intelligent water conservancy image processing method, including:
acquiring a target image, preprocessing the target image, dividing image sub-blocks, and determining the area where the image sub-blocks are located;
acquiring an image set of the same shooting area and the same shooting time of the image sub-blocks from a database, wherein the image set comprises images of at least one season;
carrying out binarization processing on the images in the image set to obtain a binarization image set;
determining a first land contour and a first water area contour according to the binarization image set;
acquiring an image to be detected, wherein the image to be detected and the image set have the same shooting area;
determining a second land contour and a second water area contour according to the binaryzation image to be detected;
calculating a ratio of an intersection to a union between the first land contour and the second land contour as a first ratio;
calculating the ratio of the intersection and union between the first water area contour and the second water area contour as a second ratio;
and when the first ratio and the second ratio meet a preset condition, determining a predicted water level, and when the predicted water level exceeds a preset water level threshold, starting an emergency plan.
Optionally, the obtaining a target image, preprocessing the target image, dividing image subblocks, and determining a region where the image subblock is located includes:
extracting hyperspectral data in the target image;
processing the hyperspectral data through a depth confidence network to obtain terrain features;
dividing the target image into at least one image sub-block with the length of a first preset pixel and the width of a second preset pixel;
and establishing a plane coordinate system by taking the boundary point of the target image as an origin, and determining the positions of the image sub-blocks.
Optionally, after establishing a plane coordinate system with the boundary point of the target image as an origin and determining the positions of the image sub-blocks, the method further includes:
inputting the topographic features corresponding to each image sub-block into at least one classifier, and judging whether the image sub-blocks have water areas;
if the water area does not exist, taking the coordinate area where the image subblock is located as a non-detection area, and setting the gray scale of the image subblock to be 255;
if the water area exists, setting the gray scale of the water area in the image sub-block to be 0, setting the gray scale of the land area to be 255, and taking the coordinate area where the image sub-block with the water area is located as the area to be detected according to the corner point coordinates, wherein the image to be detected is obtained from the range of the area to be detected.
Optionally, the binarizing processing the images in the image set to obtain a binarized image set includes:
determining coordinates of each image in the image set in the plane coordinate system, and judging whether the image corresponds to the position of the corresponding image sub-block;
if so, carrying out land and water classification on the image set to obtain a classification result, wherein the images in the same season and in the same place are classified as a group;
giving different confidence degrees to different seasons, and performing fusion calculation on the confidence degrees and the classification results to obtain recognition results, wherein the recognition results comprise water areas and land areas, the first confidence degree corresponds to winter and spring, the second confidence degree corresponds to summer and autumn, and the first confidence degree is larger than or equal to the second confidence degree;
judging whether the area where the image subblock is located completely covers the water area in the identification result;
and if the image is completely covered, carrying out binarization processing on the images in the image set to obtain the binarization image set.
Optionally, after the determining whether the area where the image sub-block is located completely covers the water area in the recognition result, the method further includes:
and if the hyperspectral image cannot be completely covered, optimizing the classifier according to the recognition result, and returning to the step of extracting the hyperspectral data in the target image.
Optionally, when the first ratio and the second ratio satisfy a preset condition, determining a predicted water level, and when the predicted water level exceeds a preset water level threshold, starting an emergency plan includes:
acquiring reference data, and judging whether the image to be detected has sending conditions or not according to the reference data, wherein the reference data comprises past year water level data, real-time meteorological data and past year meteorological data, and the sending conditions comprise that a rainfall prediction value is greater than or equal to a preset rainfall threshold value;
if the image prediction method is available, obtaining an image set to be predicted according to the image to be predicted, and determining the predicted water level according to the image set to be predicted, wherein the image set to be predicted comprises a pre-sunrise image, a post-sunrise image, a midday image, a pre-sunset image and a post-sunset image of the position where the image to be predicted is located.
Optionally, the determining the prediction horizon according to the image set to be predicted comprises:
processing the image set to be predicted through an image classification model to obtain a time-interval predicted water level;
judging whether the time interval predicted water level is in a single correlation relation with time according to the time interval corresponding to the time interval predicted water level;
if so, taking the time-interval predicted water level as the predicted water level;
if not, executing a preset strategy, wherein the preset strategy comprises informing a worker or inputting the database.
Optionally, after the calculating a ratio of an intersection and a union between the first water area contour and the second water area contour as a second ratio, the method further includes:
when the first ratio and the second ratio do not meet a preset condition, recording season information, time information and position information of the image to be detected and the image to be detected, and inputting the season information, the time information and the position information into the database, wherein the preset condition comprises that the difference between the first ratio and the second ratio is larger than or equal to a first quantity threshold.
Compared with the prior art, the target image is preprocessed, then the target image is divided into image sub-blocks, the image set is obtained according to the image sub-blocks, binarization is further performed on the image set and is used as a basic image for subsequent classification judgment, the image set and the image sub-blocks are guaranteed to have corresponding relation, the basic image is cross-verified, and the effectiveness of the data set is improved; then, processing the land contour and the water area contour obtained by the image to be detected, judging the ratio of the two contours, and ensuring that the change conditions of the water area and the land contour are accurately obtained; when the first ratio and the second ratio meet the preset condition, the change of the representation outline is large, the predicted water level of the area in the predicted image in a future period of time is further determined, when the predicted water level exceeds the preset water level threshold, the emergency plan is started, the accuracy of the predicted water level is guaranteed, when the emergency plan needs to be started, the rapid positioning of disasters can be achieved through the area corresponding to the image sub-blocks, and therefore the purpose of accurately monitoring the abnormal water level is achieved.
In another aspect, the present invention further provides an intelligent water conservancy image processing apparatus, including:
the system comprises a preprocessing module, a storage module and a display module, wherein the preprocessing module is used for acquiring a target image, preprocessing the target image, dividing image subblocks and determining the area where the image subblocks are located;
an image set acquisition module for acquiring an image set of the same shooting area and the same shooting time as the image sub-blocks in a database, wherein the image set comprises images of at least one season;
the image acquisition module to be detected is used for acquiring an image to be detected, wherein the image to be detected and the image set have the same shooting area;
the binarization processing module is used for carrying out binarization processing on the images in the image set to obtain a binarization image set, and carrying out binarization processing on the image to be detected to obtain a binarization image to be detected;
a first contour determination module for determining a first land contour and a first water area contour from the binarized image set;
the second contour determining module is used for determining a second land contour and a second water area contour according to the binaryzation image to be detected;
a first calculation module for calculating a ratio of intersection to union between the first land contour and the second land contour as a first ratio;
a second calculation module for calculating a ratio of an intersection to a union between the first land contour and the second land contour as a second ratio;
and the central calculation module is used for determining a predicted water level when the first ratio and the second ratio meet a preset condition, and starting an emergency plan when the predicted water level exceeds a preset water level threshold.
Compared with the prior art, the intelligent water conservancy image processing device has the same beneficial effects as the intelligent water conservancy image processing method, and details are not repeated here.
In a third aspect, the invention further provides an intelligent water conservancy image processing system, which includes a processing module and a computer readable storage medium, wherein the processing module is controlled by applying the intelligent water conservancy image processing method.
Compared with the prior art, the intelligent water conservancy image processing system has the same beneficial effects as the intelligent water conservancy image processing method, and is not repeated herein.
Drawings
FIG. 1 is a flowchart illustrating steps of a method for processing an intelligent water conservancy image according to an embodiment of the present invention;
FIG. 2 is a flowchart illustrating a step S100 of an intelligent water conservancy image processing method according to an embodiment of the present invention after being refined;
FIG. 3 is another flowchart illustrating a step S100 of the intelligent water conservancy image processing method according to the embodiment of the present invention after being refined;
FIG. 4 is a flowchart illustrating a step S300 of an intelligent water conservancy image processing method according to an embodiment of the present invention after being refined;
FIG. 5 is a flowchart illustrating a step S1000 of an intelligent water conservancy image processing method according to an embodiment of the present invention after refinement;
fig. 6 is a flowchart illustrating a step S1020 of the intelligent water conservancy image processing method according to the embodiment of the invention after being refined.
Detailed Description
In order to make the aforementioned objects, features and advantages of the present invention comprehensible, embodiments accompanied with figures are described in detail below. While certain embodiments of the present invention are shown in the drawings, it should be understood that the present invention may be embodied in various forms and should not be construed as limited to the embodiments set forth herein, but rather are provided for a more complete and thorough understanding of the present invention. It should be understood that the drawings and the embodiments of the present invention are illustrative only and are not intended to limit the scope of the present invention.
It should be understood that the various steps recited in method embodiments of the present invention may be performed in a different order, and/or performed in parallel. Moreover, method embodiments may include additional steps and/or omit performing the illustrated steps. The scope of the invention is not limited in this respect.
The term "include" and variations thereof as used herein are open-ended, i.e., "including but not limited to". The term "based on" is "based, at least in part, on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments"; the term "optionally" means "alternative embodiments". Relevant definitions for other terms will be given in the following description. It should be noted that the terms "first", "second", and the like in the present invention are only used for distinguishing different devices, modules or units, and are not used for limiting the order or interdependence relationship of the functions performed by the devices, modules or units.
It is noted that references to "a", "an", and "the" modifications in the present invention are intended to be illustrative rather than limiting, and that those skilled in the art will recognize that reference to "one or more" unless the context clearly dictates otherwise.
As shown in fig. 1, an embodiment of the invention provides a method for processing an intelligent water conservancy image, including:
step S100, acquiring a target image, preprocessing the target image, dividing image subblocks, and determining the region where the image subblocks are located.
Specifically, the invention has at least two edge calculation modules, wherein the first edge calculation module is used for preprocessing the target image with lower calculation amount, and comprises the steps of dividing the target image to obtain a plurality of image sub-blocks, determining the area where each image sub-block is located, and ensuring to reduce the calculation pressure of the central processing module and reduce other transmission problems caused in the transmission process through the first edge calculation module.
In one embodiment, the target image comprises a remote sensing image, and the remote sensing image has the characteristics of large detection range, instantaneous imaging and real-time transmission and can quickly acquire target area information, so that rapid mapping can be realized in areas with severe natural conditions and difficult ground work expansion, and the integrity of the target and the information of the surrounding areas is ensured. The remote sensing image of the target area is used as a target image and then transmitted to a first edge calculation module, the target image is segmented, image subblocks are divided, a specific area of the image subblocks is determined according to the area where the target image is located, it is guaranteed that for the same water area, observation and prediction can be conducted in a cross mode, data cannot be distorted due to transmission or acquisition abnormity, on the other hand, the remote sensing image can be associated with a geodetic coordinate system, and effectiveness and readability of the data are guaranteed.
The ground image referred by the invention comprises an image obtained from the atmosphere such as ground monitoring equipment, an unmanned aerial vehicle and the like.
Step S200, acquiring an image set of the same shooting time and the same shooting area with the image sub-blocks from a database, wherein the image set comprises images of at least one season.
In one embodiment, the data stored in the database includes images obtained at the surface of the same area as the sub-block of images. The acquired image set and the image sub-blocks have the same shooting time and the same shooting area, for example, the target image is set in a range from 13:00, then 13:00, ground images of the same place are shot, and an image set is used for providing information acquired from the ground and mutually supporting with a target image so as to improve robustness.
Preferably, the image set includes images of the same shooting season as the image sub-blocks. The characteristics of land areas and water areas are different due to different seasons, for example, in summer, the characteristics of the water areas can be covered by vegetation of flowers and plants; in winter, the vegetation is withered more, and water areas in certain areas can be frozen; in the flood season, the river water area may have more obvious characteristics, so that the images in the same shooting season as the image sub-blocks are obtained to ensure the consistency of the image data in the two image sub-blocks.
Specifically, the phrase "the image set and the image sub-block are captured at the same time" means that they are captured at the same time during the day, but the year, month and day of the capturing time may be different, for example, the image sub-block is captured at 6/1/12 at noon in 2020: 00, the image capture time in the image collection may be 2000, 8, 1, noon, 12:00.
in one embodiment, the image set comprises images in the same shooting season as the image sub-blocks and images in different seasons, and after the basic data sources are ensured to be consistent, the diversity of the data sources is increased by adding the images in different seasons, so that the robustness of identification is ensured to be increased.
And step S300, carrying out binarization processing on the images in the image set to obtain a binarization image set.
In an embodiment, the image set is processed by the second edge calculation module, the image is subjected to binarization processing, the gray level of a water area in the image is set to be 0, and the gray level of a land area is set to be 255, that is, the water area is turned to be black, and the land area is turned to be white, so that the calculation amount in the image identification process is simplified, and the timeliness is improved.
And S400, determining a first land contour and a first water area contour according to the binarization image set.
Specifically, the first land contour comprises a contour of a white area where land is located, the water area contour comprises a contour of a black area where water is located, and the binary image set is divided into contours for comparison with land contours and water area contours of subsequent images to be detected.
And S500, acquiring an image to be detected, wherein the image to be detected and the image set have the same shooting area.
And S600, carrying out binarization processing on the image to be detected to obtain a binarized image to be detected.
And S700, determining a second land contour and a second water area contour according to the binary image to be detected.
Specifically, acquiring an image to be detected which is the same as a shooting area of an image set, wherein the image to be detected comprises a remote sensing image and a ground image, carrying out binarization processing on the ground image to obtain an image comprising a land contour and a water area contour, and then comparing the first ground contour with the second ground contour to obtain a variation of the ground contour; and comparing the first water area profile with the second water area profile to obtain the variation of the water area profile.
Step S800, calculating a ratio of an intersection and a union between the first land contour and the second land contour as a first ratio.
Step S900, calculating a ratio of an intersection and a union between the first water area contour and the second water area contour as a second ratio.
Specifically, a first land contour in the image set is compared with a second land contour of the image to be detected, and the variation of the land contour is judged; and comparing the first water area contour with the second water area contour in the image set, judging the variation of the water area contour, and respectively obtaining a first ratio and a second ratio.
Step S1000, when the first ratio and the second ratio meet a preset condition, determining a predicted water level, and when the predicted water level exceeds a preset water level threshold, starting an emergency plan.
The change conditions of the water area outline and the land outline can be obtained through the first ratio and the second ratio, when the first ratio and the second ratio meet preset conditions, the outline change quantity is larger than the preset change quantity, namely the change area of the water area/land area is too large, the prediction is carried out through a central calculation module with higher calculation capacity, the predicted water level is obtained, and then whether an emergency plan needs to be started or not is determined according to the predicted water level.
Optionally, the preset condition includes that both the first ratio and the second ratio are greater than or equal to a preset value.
In an embodiment, the preset value is 10%, which means that the variation of the land contour in the binarized image to be detected and the land contour at the same position in the image set exceeds 10%, and the variation of the water area contour in the binarized image to be detected and the water area contour at the same position in the image set also exceeds 10%, which indicates that the variation value of the water area contour is too large, further judgment and prediction are needed, the image is sent to a central computing module for prediction, and then whether an abnormal condition occurs in an area contained in the image to be detected is determined according to the predicted water level.
Optionally, as shown in fig. 2, the acquiring a target image, preprocessing the target image, dividing image sub-blocks, and determining a region where the image sub-block is located includes:
step S110, extracting hyperspectral data in the target image;
step S120, the hyperspectral data are processed through a deep belief network, and topographic features are obtained;
step S130, dividing the target image into at least one image sub-block with the length of a first preset pixel and the width of a second preset pixel;
and step S140, establishing a plane coordinate system by taking the boundary point of the target image as an origin, and determining the positions of the image sub-blocks.
In an embodiment, the target image includes hyperspectral data, and the hyperspectral data is processed through a depth confidence network to obtain preliminarily classified terrain features, wherein the preliminarily classified terrain features include lakes, rivers, forests, lands and the like. The depth confidence network is trained in an unsupervised learning mode, so that the terrain features in the image are classified preliminarily under the condition of lacking or not having priori knowledge, the calculation pressure of the central processing module is reduced, the target image can be preprocessed through the first edge calculation module, and the real-time performance is improved. The hyperspectral data comprises richer earth surface image information, and the deep confidence network can be identified more accurately.
In another embodiment, the target image is divided into at least one image sub-block, wherein the length and the width of the image sub-block are respectively preset length and width, so that the recorded image sub-block is ensured to have a certain segmentation standard, the complexity of data recording is reduced, and the storage and identification efficiency is increased.
In another embodiment, a plane coordinate system is established by taking a boundary point of a target image as an origin to determine a specific region of an image subblock, for example, the lower left corner of the target image is taken as the origin to establish the plane coordinate system, and then region coordinates of the image subblock are described by coordinates in the coordinate system, so that the image subblock is successfully restored to the arrangement of the target image after being processed, a comprehensive regional water area image is ensured to be obtained, the positioning of the image subblock is facilitated, and the selection of the position of an image to be detected is facilitated.
Optionally, as shown in fig. 3, after establishing a plane coordinate system with the boundary point of the target image as an origin, and determining the positions of the image sub-blocks, the method further includes:
step S150, inputting the terrain features corresponding to each image sub-block into at least one classifier, and judging whether the image sub-blocks have water areas or not;
step S160, if the water area does not exist, taking the coordinate area where the image subblock is located as a non-detection area, and setting the gray scale of the image subblock to be 255;
step S170, if the water area exists, setting the gray scale of the water area in the image sub-block to be 0, setting the gray scale of the land area to be 255, and taking the coordinate area where the image sub-block with the water area is located as the area to be detected according to the corner point coordinates, wherein the image to be detected is obtained from the range of the area to be detected.
Optionally, the classifier is a binary classification model.
In one embodiment, the terrain features of the image sub-block are input into a classifier, whether the image sub-block has a water area or not is judged, if yes, the image sub-block needs to be further processed, the area represented by the image sub-block has an observed value, the image sub-block is subjected to binarization processing, the grayscale of the water area is set to be 0, the grayscale of the land area is set to be 255, namely, the water area is set to be black, and the land area is set to be white. The region defined by the angular point coordinates is used as a region to be detected, when detection is needed, an image to be detected is obtained from the region to be detected, and therefore the region in the image to be detected can be guaranteed to have a water area.
When the image subblock does not have a water area, the gray scale of the image subblock is set to be 255, that is, the whole image subblock is set to be black, and an area formed by connecting the corner points of the image subblock is used as a non-detection area, which indicates that the area does not have the water area. When detection is needed, the image to be detected is not acquired from the region, and the calculation amount is reduced. Through dividing the region that the image subblock belongs to and waiting to detect region and non-detection area, can guarantee to detect that the image of waiting all obtains from the region that has the waters, guarantee to reduce central processing unit's calculation volume.
Optionally, as shown in fig. 4, the performing binarization processing on the images in the image set to obtain a binarized image set includes:
step S310, determining the coordinates of each image in the image set in the plane coordinate system, and judging whether the image corresponds to the position of the corresponding image sub-block;
step S320, if the images correspond to each other, the image sets are classified on land and water to obtain classification results, wherein the images in the same season and in the same place are classified as a group;
step S330, giving different confidence degrees to different seasons, and performing fusion calculation on the confidence degrees and the classification results to obtain recognition results, wherein the recognition results comprise water areas and land areas, a first confidence degree is corresponding to winter and spring, a second confidence degree is corresponding to summer and autumn, and the first confidence degree is larger than or equal to the second confidence degree;
step S340, judging whether the area where the image sub-block is located completely covers the water area in the identification result;
and step S350, if the image is completely covered, carrying out binarization processing on the images in the image set to obtain the binarized image set.
In one embodiment, images are selected from the image set for coordinate confirmation, whether the acquired images correspond to the area covered by the image subblocks in the remote sensing image or not is further judged, and if the acquired images do not correspond to the area covered by the image subblocks in the remote sensing image, the images are acquired again until the acquired images correspond to the area covered by the image subblocks, so that the validity and the accuracy of the compared images are guaranteed.
If the images correspond to each other, the images in the image set are identified to obtain classification of the water areas and the land areas, and for each season, a corresponding classification result is obtained, for example, for a certain image, the images in spring are taken as a group of independent classifications, the images in summer are taken as a group of independent classifications, and a spring land classification result and a summer land classification result are obtained. Adding confidence into the classification result, wherein the result obtained by the classifier is a probability value, for example, a first pixel with a high probability of being a land is marked as 0.9, and a second pixel with a low probability of being the land is marked as 0.1 in the image, and then dividing the first pixel into the land and the second pixel into a water area; the vegetation cover in winter and spring is less, the interference is less, the land and the water marked out by the identification model are more accurate, and the image acquired in winter and spring is set with higher confidence; and vegetation in summer covers much, leaves in autumn and the like also have strong interference effects, and the images acquired in summer and autumn are set to have lower confidence. After the confidence coefficient is added, data of four seasons are fused, and the land and water classification results of the four seasons are integrated to obtain a final recognition result, so that the accuracy of the recognition result is guaranteed, and the recognition result cannot be interfered by a certain image or certain distorted images. And after the identification result is obtained, judging whether the area surrounded by the four corner points of the image subblock can completely cover the identification result, if so, indicating that the classification result of the image subblock is consistent with the identification result of the image set, and performing binarization processing on the images in the image set to obtain a binarization image set which is used as a reference for judging whether the water area changes.
Optionally, after the determining whether the area where the image sub-block is located completely covers the water area in the recognition result, the method further includes:
and if the hyperspectral image cannot be completely covered, optimizing the classifier according to the recognition result, and returning to the step of extracting the hyperspectral data in the target image.
Because the identification result of the image set is more accurate, the inaccurate classification of the depth confidence model and the classifier thereof cannot be completely covered, optimization is needed, the step of extracting the hyperspectral data in the target image is returned, and the data acquisition and classification are carried out again.
Optionally, as shown in fig. 5, when the first ratio and the second ratio satisfy a preset condition, determining a predicted water level, and when the predicted water level exceeds a preset water level threshold, starting an emergency plan includes:
step S1010, acquiring reference data, and judging whether the image to be detected has sending conditions or not according to the reference data, wherein the reference data comprises previous year water level data, real-time meteorological data and previous year meteorological data, and the sending conditions comprise that a rainfall prediction value is greater than or equal to a preset rainfall threshold value;
and step S1020, if the image to be predicted is available, obtaining an image set to be predicted according to the image to be predicted, determining the predicted water level according to the image set to be predicted, wherein the image set to be predicted comprises a pre-sunrise image, a post-sunrise image, a noon image, a pre-sunset image and a post-sunset image of the position of the image to be predicted, and determining the predicted water level according to the image set to be predicted.
Optionally, after the calculating a ratio of an intersection and a union between the first water area contour and the second water area contour as a second ratio, the method further includes:
when the first ratio and the second ratio do not meet a preset condition, recording season information, time information and position information of the image to be detected and the image to be detected, and inputting the season information, the time information and the position information into the database, wherein the preset condition comprises that the difference between the first ratio and the second ratio is larger than or equal to a first quantity threshold.
In an embodiment, when the difference between the first ratio and the second ratio is greater than or equal to the first quantity threshold, it indicates that the change between the first ratio and the second ratio is large, that is, the change between the first land contour and the second land contour and the change between the first water area contour and the second water area contour are large, and further prediction is required to determine whether a special situation occurs. Therefore, the past year water level data, the weather data and the real-time weather data are used as references to judge whether a weather disaster or other special conditions occur in the time period in the past year, and if the transmission conditions are not met, namely similar conditions exist in the past year and no disaster occurs, the data are uploaded to the database and stored as historical data.
If the image to be predicted has the sending condition, the disaster event is possible to happen, the image set to be predicted is sent to the central computing module, whether the natural disaster happens or not is predicted in a combined mode according to the five images of the image before sunrise, the image after sunrise, the image at noon, the image before sunset and the image after sunset, for example, when the image before sunrise and the image at noon can predict the natural disaster, the other three images are in normal conditions, and the result is notified to workers; and when the natural disaster occurs with probability in the result obtained by all the images, obtaining the predicted water level based on the water level trends of the five images.
Optionally, as shown in fig. 6, the determining the predicted water level according to the image set to be predicted includes:
step S1021, processing the image set to be predicted through an image classification model to obtain a time-interval prediction water level;
step S1022, judging whether the time interval predicted water level is in a single correlation relation with time according to the time interval corresponding to the time interval predicted water level;
step S1023, if yes, taking the time interval predicted water level as the predicted water level;
and step S1024, if not, executing a preset strategy, wherein the preset strategy comprises informing a worker or entering the database.
Preferably, the ResNet18 or ResNet34 model is trained to obtain an image classification model.
And processing the image set to be predicted according to the image classification model to obtain predicted water levels in different time periods, and then judging whether the predicted water levels in different time periods have a single correlation relationship, such as the water level increases or decreases along with the time. If the single correlation exists, the time-interval predicted water level is used as the predicted water level, and when the predicted water level is higher than a preset water level threshold, an emergency plan is started.
And when the single correlation relation does not exist, further judging whether the predicted highest water level is higher than a preset water level threshold value, if so, starting an emergency plan, otherwise, if not, not definitely judging whether a disaster occurs, and executing a preset strategy.
In another aspect, an embodiment of the present invention provides an intelligent water conservancy image processing apparatus, including:
the system comprises a preprocessing module, a storage module and a display module, wherein the preprocessing module is used for acquiring a target image, preprocessing the target image, dividing image subblocks and determining the area where the image subblocks are located;
an image set acquisition module for acquiring an image set of the same shooting area and the same shooting time as the image sub-blocks in a database, wherein the image set comprises images of at least one season;
the image acquisition module to be detected is used for acquiring an image to be detected, wherein the image to be detected and the image set have the same shooting area;
the binarization processing module is used for carrying out binarization processing on the images in the image set to obtain a binarization image set and carrying out binarization processing on the image to be detected to obtain a binarization image to be detected;
a first contour determination module for determining a first land contour and a first water area contour from the binarized image set;
the second contour determining module is used for determining a second land contour and a second water area contour according to the binaryzation image to be detected;
a first calculation module for calculating a ratio of intersection to union between the first land contour and the second land contour as a first ratio;
a second calculation module for calculating a ratio of an intersection to a union between the first and second land contours as a second ratio;
and the central calculation module is used for determining a predicted water level when the first ratio and the second ratio meet a preset condition, and starting an emergency plan when the predicted water level exceeds a preset water level threshold.
Specifically, at least one edge calculation module is used for performing operation tasks except that the step of determining the predicted water level when the first ratio and the second ratio meet a preset condition and starting an emergency plan when the predicted water level exceeds a preset water level threshold is performed.
In one embodiment, the preprocessing module, the image set acquisition module, the image acquisition module to be detected, the binarization processing module, the first contour determination module, the second contour determination module, the first calculation module and the second calculation module are all independently calculated by respective independent edge calculation devices, and are used for sharing the calculation pressure of the central calculation module, and further increasing the water level prediction speed of the central calculation module.
In a third aspect, an embodiment of the present invention provides an intelligent water conservancy image processing system, including a processing module and a computer readable storage medium, where the processing module is controlled by applying the intelligent water conservancy image processing method described above.
Compared with the prior art, the intelligent water conservancy image processing system has the same beneficial effects as the intelligent water conservancy image processing method, and details are not repeated here.
Another embodiment of the present invention provides an electronic device, including a memory and a processor; the memory for storing a computer program; the processor is used for realizing the intelligent water conservancy image processing method when the computer program is executed.
A further embodiment of the present invention provides a computer-readable storage medium, which stores a computer program, and when the computer program is executed by a processor, the method for processing intelligent water conservancy images as described above is implemented.
An electronic device that may be a server or a client of the present invention, which is an example of a hardware device that may be applied to aspects of the present invention, will now be described. Electronic device is intended to represent various forms of digital electronic computer devices, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframes, and other appropriate computers. The electronic device may also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions, are meant to be exemplary only, and are not meant to limit implementations of the inventions described and/or claimed herein.
The electronic device includes a computing unit that can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) or a computer program loaded from a storage unit into a Random Access Memory (RAM). In the RAM, various programs and data required for the operation of the device can also be stored. The computing unit, the ROM, and the RAM are connected to each other by a bus. An input/output (I/O) interface is also connected to the bus.
The computer system may include clients and servers. A client and server are generally remote from each other and typically interact through a communication network. The relationship of client and server arises by virtue of computer programs running on the respective computers and having a client-server relationship to each other.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by a computer program, which can be stored in a computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. The storage medium may be a magnetic disk, an optical disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), or the like. In this application, the units described as separate parts may or may not be physically separate, and parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of the embodiments of the present invention. In addition, functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit. The integrated unit can be realized in a form of hardware, and can also be realized in a form of a software functional unit.
Although the present disclosure has been described above, the scope of the present disclosure is not limited thereto. Various changes and modifications may be effected therein by one of ordinary skill in the pertinent art without departing from the spirit and scope of the present disclosure, and these changes and modifications are intended to be within the scope of the present disclosure.
Claims (10)
1. An intelligent water conservancy image processing method is characterized by comprising the following steps:
acquiring a target image, preprocessing the target image, dividing image sub-blocks, and determining the area where the image sub-blocks are located;
acquiring an image set of the same shooting time and the same shooting area as the image sub-blocks in a database, wherein the image set comprises images of at least one season;
carrying out binarization processing on the images in the image set to obtain a binarization image set;
determining a first land contour and a first water area contour according to the binarization image set;
acquiring an image to be detected, wherein the image to be detected and the image set have the same shooting area;
carrying out binarization processing on the image to be detected to obtain a binarized image to be detected;
determining a second land profile and a second water area profile according to the binaryzation image to be detected;
calculating a ratio of an intersection to a union between the first land contour and the second land contour as a first ratio;
calculating the ratio of the intersection and union between the first water area contour and the second water area contour as a second ratio;
and when the first ratio and the second ratio meet a preset condition, determining a predicted water level, and when the predicted water level exceeds a preset water level threshold, starting an emergency plan.
2. The intelligent water conservancy image processing method according to claim 1, wherein the acquiring a target image, preprocessing the target image, dividing image sub-blocks, and determining the region where the image sub-blocks are located comprises:
extracting hyperspectral data in the target image;
processing the hyperspectral data through a depth confidence network to obtain terrain features;
dividing the target image into at least one image sub-block with the length of a first preset pixel and the width of a second preset pixel;
and establishing a plane coordinate system by taking the boundary point of the target image as an origin, and determining the positions of the image sub-blocks.
3. The intelligent water conservancy image processing method according to claim 2, wherein after establishing a plane coordinate system with a boundary point of the target image as an origin and determining the positions of the image sub-blocks, the method further comprises:
inputting the topographic features corresponding to each image sub-block into at least one classifier, and judging whether the image sub-blocks have water areas;
if the water area does not exist, taking the coordinate area where the image subblock is located as a non-detection area, and setting the gray scale of the image subblock to be 255;
if the water area exists, setting the gray scale of the water area in the image sub-block to be 0, setting the gray scale of the land area to be 255, and taking the coordinate area where the image sub-block with the water area is located as the area to be detected according to the corner point coordinates, wherein the image to be detected is obtained from the range of the area to be detected.
4. The intelligent water conservancy image processing method according to claim 3, wherein the binarizing processing of the images in the image set to obtain a binarized image set comprises:
determining coordinates of each image in the image set in the plane coordinate system, and judging whether the image corresponds to the position of the corresponding image sub-block;
if so, carrying out land and water classification on the image set to obtain a classification result, wherein the images in the same season and in the same place are classified as a group;
giving different confidence degrees to different seasons, and performing fusion calculation on the confidence degrees and the classification results to obtain recognition results, wherein the recognition results comprise water areas and land areas, the winter and the spring correspond to a first confidence degree, the summer and the autumn correspond to a second confidence degree, and the first confidence degree is larger than or equal to the second confidence degree;
judging whether the area where the image subblock is located completely covers the water area in the identification result or not;
and if the images are completely covered, carrying out binarization processing on the images in the image set to obtain the binarization image set.
5. The method of claim 4, wherein after determining whether the area where the image sub-block is located completely covers the water area in the recognition result, the method further comprises:
and if the hyperspectral data cannot be completely covered, optimizing the classifier according to the recognition result, and returning to the step of extracting the hyperspectral data in the target image.
6. The intelligent water conservancy image processing method according to claim 5, wherein the determining a predicted water level when the first ratio and the second ratio satisfy a preset condition comprises:
acquiring reference data, and judging whether the image to be detected has sending conditions or not according to the reference data, wherein the reference data comprises past year water level data, real-time meteorological data and past year meteorological data, and the sending conditions comprise that a rainfall prediction value is greater than or equal to a preset rainfall threshold value;
if the image prediction method is available, obtaining an image set to be predicted according to the image to be predicted, and determining the predicted water level according to the image set to be predicted, wherein the image set to be predicted comprises a pre-sunrise image, a post-sunrise image, a midday image, a pre-sunset image and a post-sunset image of the position where the image to be predicted is located.
7. The method of claim 6, wherein said determining the predicted water level according to the image set to be predicted comprises:
processing the image set to be predicted through an image classification model to obtain a time-interval predicted water level;
judging whether the time interval predicted water level is in a single correlation relation with time according to the time interval corresponding to the time interval predicted water level;
if so, taking the time-interval predicted water level as the predicted water level;
and if not, executing a preset strategy, wherein the preset strategy comprises informing a worker or entering the database.
8. The method according to any one of claims 1-7, further comprising, after said calculating a ratio of intersection to union between said first water area contour and said second water area contour as a second ratio:
when the first ratio and the second ratio do not meet a preset condition, recording season information, time information and position information of the image to be detected and the image to be detected, and inputting the season information, the time information and the position information into the database, wherein the preset condition comprises that the difference between the first ratio and the second ratio is larger than or equal to a first quantity threshold.
9. An intelligent water conservancy image processing device is characterized by comprising:
the system comprises a preprocessing module, a storage module and a display module, wherein the preprocessing module is used for acquiring a target image, preprocessing the target image, dividing image subblocks and determining the area where the image subblocks are located;
an image set acquisition module for acquiring an image set of the same shooting region and the same shooting time in a database as the image sub-blocks, wherein the image set comprises images of at least one season;
the image acquisition module to be detected is used for acquiring an image to be detected, wherein the image to be detected and the image set have the same shooting area;
the binarization processing module is used for carrying out binarization processing on the images in the image set to obtain a binarization image set and carrying out binarization processing on the image to be detected to obtain a binarization image to be detected;
a first contour determination module for determining a first land contour and a first water area contour from the binarized image set;
the second contour determining module is used for determining a second land contour and a second water area contour according to the binaryzation image to be detected;
a first calculation module for calculating a ratio of intersection to union between the first land contour and the second land contour as a first ratio;
a second calculation module for calculating a ratio of an intersection to a union between the first and second land contours as a second ratio;
and the central calculation module is used for determining a predicted water level when the first ratio and the second ratio meet a preset condition, and starting an emergency plan when the predicted water level exceeds a preset water level threshold.
10. An intelligent water conservancy image processing system, which comprises a processing module and a computer readable storage medium, wherein the processing module is controlled by applying the intelligent water conservancy image processing method according to any one of claims 1-8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211676362.3A CN115965639B (en) | 2022-12-26 | 2022-12-26 | Intelligent water conservancy image processing method, device and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211676362.3A CN115965639B (en) | 2022-12-26 | 2022-12-26 | Intelligent water conservancy image processing method, device and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115965639A true CN115965639A (en) | 2023-04-14 |
| CN115965639B CN115965639B (en) | 2023-08-29 |
Family
ID=87361042
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211676362.3A Active CN115965639B (en) | 2022-12-26 | 2022-12-26 | Intelligent water conservancy image processing method, device and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115965639B (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20200250394A1 (en) * | 2019-01-31 | 2020-08-06 | Palantir Technologies Inc. | Systems and methods for coherent monitoring |
| WO2021238030A1 (en) * | 2020-05-26 | 2021-12-02 | 浙江大学 | Water level monitoring method for performing scale recognition on the basis of partitioning by clustering |
| US20210374466A1 (en) * | 2020-05-26 | 2021-12-02 | Zhejiang University | Water level monitoring method based on cluster partition and scale recognition |
| CN114037912A (en) * | 2022-01-07 | 2022-02-11 | 成都国星宇航科技有限公司 | Method and device for detecting change of remote sensing image and computer readable storage medium |
| CN114332870A (en) * | 2021-12-31 | 2022-04-12 | 武汉大学 | Water level identification method, device, equipment and readable storage medium |
| CN114494695A (en) * | 2022-01-13 | 2022-05-13 | 广州数鹏通科技有限公司 | Intelligent water conservancy urban and rural waterlogging level monitoring and early warning method and device |
-
2022
- 2022-12-26 CN CN202211676362.3A patent/CN115965639B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20200250394A1 (en) * | 2019-01-31 | 2020-08-06 | Palantir Technologies Inc. | Systems and methods for coherent monitoring |
| WO2021238030A1 (en) * | 2020-05-26 | 2021-12-02 | 浙江大学 | Water level monitoring method for performing scale recognition on the basis of partitioning by clustering |
| US20210374466A1 (en) * | 2020-05-26 | 2021-12-02 | Zhejiang University | Water level monitoring method based on cluster partition and scale recognition |
| CN114332870A (en) * | 2021-12-31 | 2022-04-12 | 武汉大学 | Water level identification method, device, equipment and readable storage medium |
| CN114037912A (en) * | 2022-01-07 | 2022-02-11 | 成都国星宇航科技有限公司 | Method and device for detecting change of remote sensing image and computer readable storage medium |
| CN114494695A (en) * | 2022-01-13 | 2022-05-13 | 广州数鹏通科技有限公司 | Intelligent water conservancy urban and rural waterlogging level monitoring and early warning method and device |
Non-Patent Citations (5)
| Title |
|---|
| EDWAOD SMITH: "Multi-View Silhouette and Depth Decomposition for High Resolution 3D Object Representation", 32ND CONFERENCE ON NEURAL INFORMATION PROCESSING SYSTEMS * |
| KUKIL: "Intersection over Union(IoU) in object Decection & Segmentation", LEARNOPENCV * |
| OLEKSII SHEREMET: "Intersection over unio(IoU) calculation for evaluating an image segmentation model", TOWARDS DATA SCIENCE * |
| 李昕悦等: "适用于城市洪水的水体提取方法对比与分析", 地下水, no. 05 * |
| 许淑淑: "基于对象的多源数据变化检测的方法", 测绘通报, no. 1 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115965639B (en) | 2023-08-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110097536B (en) | Hexagonal bolt looseness detection method based on deep learning and Hough transform | |
| Jiao et al. | Multi-order landscape expansion index: Characterizing urban expansion dynamics | |
| Martinis et al. | Unsupervised extraction of flood-induced backscatter changes in SAR data using Markov image modeling on irregular graphs | |
| Schwegmann et al. | Manifold adaptation for constant false alarm rate ship detection in South African oceans | |
| CN106650812B (en) | A kind of urban water-body extracting method of satellite remote-sensing image | |
| CN112102288B (en) | Water body identification and water body change detection method, device, equipment and medium | |
| US10685443B2 (en) | Cloud detection using images | |
| CN104063711A (en) | Corridor vanishing point rapid detection algorithm based on K-means method | |
| CN114564545A (en) | System and method for extracting ship experience course based on AIS historical data | |
| Wan et al. | Automatic extraction of flood inundation areas from SAR images: A case study of Jilin, China during the 2017 flood disaster | |
| CN112883850A (en) | Multi-view aerospace remote sensing image matching method based on convolutional neural network | |
| KR20180020421A (en) | Method and system for extracting coastline based on a large-scale high-resolution satellite images | |
| CN113963314A (en) | Rainfall monitoring method and device, computer equipment and storage medium | |
| CN110636248B (en) | Target tracking method and device | |
| CN109241893B (en) | Road selection method and device based on artificial intelligence technology and readable storage medium | |
| CN113870224A (en) | Flood monitoring method, system, equipment and medium | |
| CN117291936A (en) | Point cloud segmentation method, device, equipment and medium | |
| CN115965639B (en) | Intelligent water conservancy image processing method, device and system | |
| CN116843946A (en) | Tunnel rock mass main structural surface identification method and device based on image identification | |
| CN114742849B (en) | Leveling instrument distance measuring method based on image enhancement | |
| CN111476129A (en) | Soil impurity detection method based on deep learning | |
| CN116630814B (en) | Quick positioning and evaluating method for building disasters based on machine learning | |
| CN117409329B (en) | Method and system for reducing false alarm rate of underground cavity detection by three-dimensional ground penetrating radar | |
| CN117765395A (en) | Road extraction method and system for generalizing semantic segmentation algorithm | |
| CN114663790B (en) | Intelligent remote sensing mapping 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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |