An extraction method, device, equipment and storage medium for a coastal zone culture pond

Abstract

The embodiment of the present application relates to the technical field of remote sensing image recognition, and extracts the method, device, equipment and storage medium of the coastal zone culture pond. The extraction method comprises steps S1 to S7. S1 acquires a to-be-recognized remote sensing image. S2 acquires coastline data of a target region, and acquires a first potential distribution range according to the coastline data. S3 performs water body and non-water body segmentation on the to-be-recognized remote sensing image in the first potential distribution range, and acquires a second potential distribution range. S4 acquires a plurality of historical remote sensing images of the second potential distribution range, and performs cloud removal processing. S5 calculates the water body frequency of each pixel according to the plurality of historical remote sensing images after the cloud removal processing, and acquires a third potential distribution range according to the water body frequency. S6 adopts a multi-scale segmentation algorithm to segment the to-be-recognized remote sensing image in the third potential distribution range, and acquires a plurality of segmentation objects. S7 adopts an object-oriented classification algorithm to recognize each segmentation object, so as to acquire a culture pond region.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of remote sensing image recognition, in particular to a method and device for extracting a coastal zone aquaculture pond, equipment and a storage medium. BACKGROUND

[0002] As a key interface of land-sea integration, the coastal zone plays an important role in addressing climate change and is an important spatial leverage for ecological environment optimization. With the booming development of aquaculture in the coastal zone, large areas of coastal wetlands have been transformed into aquaculture ponds, greatly squeezing the ecological niche of coastal wetland vegetation. A comprehensive understanding of the spatial distribution and development scale of coastal zone aquaculture ponds is crucial for the scientific management and sustainable development of coastal areas.

[0003] Traditional monitoring of coastal aquaculture ponds mostly uses field investigation methods, which are time-consuming and labor-intensive and cannot be carried out in large-scale areas. With the maturity of remote sensing image technology, techniques for extracting aquaculture ponds using optical, radar, and other multi-source remote sensing images have been developed. However, in current extraction techniques for aquaculture ponds, aquaculture ponds are easily confused with abandoned ponds, rice fields, water diversion channels, seawalls, irrigation channels, or salt ponds, which greatly limits the extraction accuracy and makes it impossible to accurately assess the spatiotemporal pattern changes of aquaculture ponds.

[0004] In view of this, the applicant has developed the present application after studying the existing technology. SUMMARY

[0005] The present application provides a method and device for extracting a coastal zone aquaculture pond, equipment and a storage medium to improve at least one of the above technical problems.

[0006] In a first aspect,

[0007] The present application provides a method for extracting a coastal zone aquaculture pond, which comprises steps S1 to S7.

[0008] S1, obtaining a remote sensing image to be recognized.

[0009] S2, obtaining coastline data of a target area and obtaining a first potential distribution range based on the coastline data.

[0010] S3, performing water body and non-water body segmentation on the remote sensing image to be recognized within the first potential distribution range to obtain a second potential distribution range.

[0011] S4, obtaining a plurality of historical remote sensing images of the second potential distribution range and performing cloud removal processing.

[0012] S5, calculating the water body frequency of each pixel based on the plurality of historical remote sensing images after cloud removal processing, and obtaining a third potential distribution range based on the water body frequency.

[0013] S6, a multi-scale segmentation algorithm is used to segment the to-be-identified remote sensing image in the third potential distribution range, and a plurality of segmentation objects are obtained.

[0014] S7, an object-oriented classification algorithm is used to identify each segmentation object to obtain the aquaculture pond area.

[0015] The second aspect,

[0016] The embodiment of the application provides an extraction device for a coastal aquaculture pond, which comprises:

[0017] An identification image acquisition module is configured to acquire a to-be-identified remote sensing image.

[0018] A first range acquisition module is configured to acquire coastline data of a target area and acquire a first potential distribution range according to the coastline data.

[0019] A second range acquisition module is configured to perform water body and non-water body segmentation on the to-be-identified remote sensing image in the first potential distribution range and acquire a second potential distribution range.

[0020] A historical image acquisition module is configured to acquire a plurality of historical remote sensing images of the second potential distribution range and perform cloud removal processing.

[0021] A third range acquisition module is configured to calculate water body frequencies of each pixel according to the plurality of historical remote sensing images after the cloud removal processing and acquire a third potential distribution range according to the water body frequencies.

[0022] An identification image segmentation module is configured to use a multi-scale segmentation algorithm to segment the to-be-identified remote sensing image in the third potential distribution range and obtain a plurality of segmentation objects.

[0023] An object classification module is configured to use an object-oriented classification algorithm to identify each segmentation object to obtain an aquaculture pond area.

[0024] The third aspect,

[0025] The embodiment of the application provides an extraction device for a coastal aquaculture pond, which comprises a processor, a memory, and a computer program stored in the memory. The computer program can be executed by the processor to implement the extraction method for the coastal aquaculture pond as described in any one of the first aspect.

[0026] The fourth aspect,

[0027] The embodiment of the present application provides a computer readable storage medium. The computer readable storage medium comprises a stored computer program, wherein the computer readable storage medium controls the device where the computer readable storage medium is located to execute the extraction method of the coastal zone culture pond as described in any one of the first aspect when the computer program runs.

[0028] By adopting the technical scheme, the present application can achieve the following technical effects:

[0029] The extraction method of the coastal zone culture pond provided by the embodiment of the present application further accurately determines the possible area of the culture pond by the water body frequency, effectively eliminates the influence of the temporary water body on the identification result, improves the identification accuracy of the culture pond, and has good practical significance. BRIEF DESCRIPTION OF DRAWINGS

[0030] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed to be used in the embodiments will be briefly introduced as follows. It should be understood that the following drawings only show some embodiments of the present application, and therefore should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can also be obtained without creative labor on the basis of these drawings.

[0031] Figure 1 is a flowchart of the extraction method.

[0032] Figure 2 is a logic block diagram of the extraction method.

[0033] Figure 3 is a conceptual diagram of water body frequency statistics.

[0034] Figure 4 is an effect diagram of using Otsu method to segment NDWI, mNDWI and AWEIsh index.

[0035] Figure 5 is a satellite image distribution trajectory of Chinese coastal zone and water body frequency statistics result.

[0036] Figure 6 is a structural schematic diagram of the extraction device. DETAILED DESCRIPTION

[0037] The technical solutions in the embodiments of the present application will be described clearly and completely in combination with the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0038] Embodiment one,

[0039] Please refer to Figures 1 to 5 The first embodiment of the present application provides an extraction method of a coastal aquaculture pond, which can be executed by an extraction device of the coastal aquaculture pond (hereinafter referred to as: extraction device). In particular, it is executed by one or more processors in the extraction device to achieve steps S1 to S7.

[0040] S1, obtaining a remote sensing image to be identified.

[0041] Specifically, the remote sensing image is a multi-band scanning image obtained by a landsat. The remote sensing image in the embodiment of the present application is surface reflectance (SR) data after radiation calibration, geometric correction and atmospheric correction. The remote sensing image to be identified is a cloud-free remote sensing image synthesized by several remote sensing images with similar time.

[0042] It can be understood that the extraction device can be a portable notebook computer, a desktop computer, a server, a smart phone or a tablet computer, etc. electronic device with computing performance.

[0043] S2, obtaining coastline data of the target area, and obtaining a first potential distribution range according to the coastline data.

[0044] Specifically, the technical problem to be solved by the present application is to identify the aquaculture pond near the coastal zone. The initial data range is obtained by the coastal zone, which can greatly reduce the amount of subsequent steps.

[0045] On the basis of the above-mentioned embodiment, in an optional embodiment of the present application, step S2 comprises steps S21 to S24.

[0046] S21, obtaining Open street map global coastline data.

[0047] S22, according to the Open street map global coastline data, extracting the region with a first preset distance on both sides of the coastline in the target area, and obtaining a buffer area. Preferably, the first preset distance is 30km.

[0048] S23, obtaining SRTM V3 digital elevation model.

[0049] S24, according to the SRTM V3 digital elevation model, extracting the region with an elevation less than a first preset height in the buffer area, and obtaining a first potential distribution range. Preferably, the first preset height is 10m.

[0050] In the embodiment, first, the global coastline data of Open street map publicly available is used to generate a 30km buffer zone inside and outside the coastline, and the potential distribution range of the coastal aquaculture pond is preliminarily obtained. Then, the flat open area with an elevation less than 10m is determined as the first potential distribution range of the coastal aquaculture pond according to the SRTM V3 (SRTM Plus) data set provided by the Google Earth Engine cloud computing platform.

[0051] Specifically, the global coastline data of Open street map is publicly available data that can be directly obtained. The SRTM V3 digital elevation model is provided by the Google Earth Engine cloud computing platform and is also a directly available elevation model.

[0052] The initial range on both sides of the coastline is obtained through the coastline data, which can effectively obtain the initial area. It can be understood that the elevation of the aquaculture pond near the coastal zone is mostly close to sea level. In the embodiment, the buffer zone area is further screened by elevation, so as to obtain a more accurate potential distribution range.

[0053] S3, the water body and non-water body segmentation is performed on the to-be-identified remote sensing image in the first potential distribution range, and the second potential distribution range is obtained.

[0054] Specifically, different ground surfaces reflect different light, and through the band data of the remote sensing image, the ground surface can be identified as water body or non-water body.

[0055] On the basis of the above embodiment, in an optional embodiment of the present application, step S3 comprises steps S31 to S32.

[0056] S31, according to the to-be-identified remote sensing image in the first potential distribution range, the segmentation index of each pixel is calculated. The segmentation index is also called water body index, which is an index selected in advance to distinguish water body and non-water body.

[0057] S32, according to the segmentation index of each pixel, the water body and non-water body segmentation is performed by using a global threshold adaptive algorithm, and the second potential distribution range of the water body region is obtained.

[0058] In the embodiment, the global threshold adaptive algorithm is the Otsu threshold method. The threshold value of segmentation can be automatically determined by the Otsu threshold method, and each pixel can be classified according to the index directly under the condition of threshold value determination. The Otsu threshold method is prior art, and the present application will not be described here.

[0059] On the basis of the above embodiment, in an optional embodiment of the present application, the selection step of the segmentation index is:

[0060] obtaining a remote sensing image of a target region and a standard distribution map of aquaculture ponds.

[0061] calculating a normalized difference water index, a modified normalized difference water index and an automated water extraction index of the remote sensing image of the target region, respectively.

[0062] segmenting the remote sensing image of the target region according to the three indexes respectively by using a global threshold adaptive algorithm to obtain the distribution map of aquaculture ponds of the target region corresponding to the three indexes.

[0063] comparing the distribution map of aquaculture ponds of the target region corresponding to the three indexes with the standard distribution map of aquaculture ponds, and selecting the best segmentation effect among the three indexes as the segmentation index.

[0064] Specifically, the normalized difference water index (NDWI), the modified normalized difference water index (mNDWI) and the AWEIsh index of the automated water extraction index (AWEI) are calculated according to the remote sensing image of the target region. Then, the global threshold adaptive algorithm (Otsu threshold method) is used to test the water extraction effect, and the water extraction effect is as shown in Figure 4

[0065] The first row and the first column of the mNDWI index image are shown in the first row and the first column of the mNDWI index image. Figure 4

[0066] The second row and the first column of the mNDWI index corresponding to the extraction effect are shown in the second row and the first column of the mNDWI index corresponding to the extraction effect. Figure 4

[0067] The first row and the second column of the NDWI index image are shown in the first row and the second column of the NDWI index image. Figure 4

[0068] The second row and the second column of the NDWI index corresponding to the extraction effect are shown in the second row and the second column of the NDWI index corresponding to the extraction effect. Figure 4

[0069] The first row and the third column of the AWEI index image are shown in the first row and the third column of the AWEI index image. Figure 4

[0070] The second row and the third column of the AWEI index corresponding to the extraction effect are shown in the second row and the third column of the AWEI index corresponding to the extraction effect. Figure 4

[0071] The third row and the first column of the false color image and the true color image are shown in the third row and the first column of the false color image and the true color image. Figure 4

[0072] ​In the embodiment, the best mNDWI index image is selected to segment the coastal water body and non-water body, and the water body region is extracted as the second potential distribution range. The segmentation index in the following is the modify Normalized Difference Water Index (mNDWI). In other embodiments, other existing indexes can be selected to distinguish water body and non-water body, and the specific type of the segmentation index is not limited in the present application, as long as it can be used to distinguish water body and non-water body.

[0073] S4, a plurality of historical remote sensing images of the second potential distribution range are acquired, and cloud removal processing is performed.

[0074] Specifically, due to the shielding of the cloud layer or the shadow of the cloud layer on the ground surface, the light reflected by the ground surface changes, and this part of the pixel cannot be used for the judgment of water body and non-water body, so cloud removal processing is needed to remove this part of the pixel.

[0075] On the basis of the above-mentioned embodiments, in an optional embodiment of the present application, step S4 comprises steps S41 to S42.

[0076] S41, a plurality of historical remote sensing images of the second potential distribution range in the past year are acquired, and radiation calibration, geometric correction and atmospheric correction are performed to obtain a plurality of ground surface reflectance data.

[0077] S42, cloud removal processing is performed on each ground surface reflectance data through the Google Earth Engine cloud computing platform.

[0078] Specifically, cloud removal processing is performed on each image in the buffer zone based on the Google Earth Engine cloud computing platform, and the CFmask band is used as a quality control band to maximize the prevention of the interference of cloud and cloud shadow pixels on the present research, and the remaining pixels are considered as good observation values that can be used for detection. Specifically, the cloud removal processing of the Google Earth Engine cloud computing platform is prior art, and the specific operation thereof will not be described herein.

[0079] S5, water body frequency of each pixel is calculated according to the plurality of historical remote sensing images after the cloud removal processing, and the third potential distribution range is obtained according to the water body frequency.

[0080] Specifically, the aquaculture pond and other water body facilities are all water bodies, and have similar characteristics in the spectral reflectance of satellite images. Therefore, the water body after excluding the non-water body needs to be further introduced into additional characteristics to distinguish the aquaculture pond from other water body facilities.

[0081] The inventors find that, according to the fact that the coastal aquaculture area has a higher water body frequency than the temporary water bodies such as water diversion channels, seawall rivers and irrigation channels, the embodiment of the application innovatively introduces the water body frequency information to distinguish the aquaculture ponds from other water body facilities, so as to exclude the interference of the temporary water bodies.

[0082] On the basis of the above embodiment, in an optional embodiment of the application, the step S5 comprises steps S51 to S54.

[0083] S51, obtaining the effective observation number of each pixel according to the plurality of historical remote sensing images after cloud removal.

[0084] Specifically, after filtering the pixels disturbed by clouds and cloud shadows, the cloud-removed images of different periods are stacked together at the pixel level, and the total effective observation number observed at each pixel position is counted, as shown in the formula. Figure 3

[0085] S52, calculating the segmentation index of each surface reflectance data after cloud removal, and performing water body and non-water body segmentation by using a global threshold adaptive algorithm to obtain the water body occurrence number of each pixel.

[0086] Specifically, the mNDWI index images of each historical remote sensing image are calculated, and then the water body-non-water body binary images segmented by using the global threshold adaptive algorithm are stacked together at the pixel level, and the total number of water body occurrences observed at each pixel position is counted.

[0087] S53, obtaining a water body frequency layer according to the water body occurrence number and the effective observation number.

[0088] Specifically, the total number of water body occurrences observed at each pixel position is divided by the total effective observation number of the position, and finally a water body frequency layer is generated. The formula is as follows:

[0089] FWater = ∑NWater / ∑NGood

[0090] In the formula, FWater represents the annual average surface water body pixel occurrence frequency, ∑NWater represents the total number of surface water body pixel occurrences, and ∑NGood is the total effective observation number of each pixel position.

[0091] S54, taking the water body occurrence number of the pond sample point as a threshold, obtaining a third potential distribution range according to the water body frequency layer.

[0092] Specifically, satellite images in which aquaculture ponds are concentrated are selected, sample points of aquaculture ponds and non-aquaculture ponds are established, and a water body frequency distribution histogram of the aquaculture pond sample points is counted (as shown in the formula). Figure 5 ​determining a threshold of water body frequency of the aquaculture pond. Water bodies with a water body frequency greater than the threshold of water body frequency of the aquaculture pond are determined as potential water bodies of the aquaculture pond. The precise range of the potential distribution area of the coastal aquaculture pond is further optimized.

[0093] S6, using a multi-scale segmentation algorithm, segmenting the to-be-identified remote sensing image in the third potential distribution range to obtain a plurality of segmentation objects.

[0094] Specifically, the multi-scale segmentation algorithm is used to segment the multi-band image. In the image layer weight setting stage, the Landsat image composed of a plurality of raster layers is input as the segmentation object of the scale segmentation algorithm, the same weight 1 is assigned to them at each level, the segmentation is iteratively performed by the estimation of scale parameter (ESP) method with a fixed step, and the local variance of each scale is calculated to realize fast and objective parameterization, and finally the best segmentation scale, shape parameter value and tightness weight value are determined.

[0095] S7, using an object-oriented classification algorithm, identifying each segmentation object to obtain the aquaculture pond area.

[0096] Specifically, the spectral information and water body frequency of the water body of the coastal aquaculture pond and the water body such as rivers, lakes and offshore waters are highly similar, and cannot be distinguished by the global threshold adaptive method and the water body seasonal fluctuation information. Considering that the water body of the aquaculture pond has a unique spatial structure feature (rectangle and square, etc.) compared with the water body such as rivers, lakes and offshore waters, the object-oriented classification method is used to effectively distinguish such ground objects,

[0097] In this embodiment, the remote sensing image is first segmented into block objects, and then the object-oriented classification algorithm is used to identify the block objects, so that the aquaculture pond can be accurately identified.

[0098] On the basis of the above embodiment, in an optional embodiment of the present application, step S7 comprises steps S71 to S72.

[0099] S71, selecting a feature space by a feature space optimization method. The feature space includes the mean values of bands 1, 4, 5, 6 and 8, the modified normalized difference water index, the water body frequency and the set feature.

[0100] S72, according to the feature space of the segmentation object, classifying each object as an aquaculture pond or a non-aquaculture pond by a nearest neighbor classifier to obtain the aquaculture pond area.

[0101] Specifically, referring to Google Earth satellite images and visual interpretation, a coastal aquaculture pond sample dataset and a non-aquaculture pond area sample dataset (including nearshore waters, rivers, and lakes) are established, and the best feature space is optimized by a Feature Space Optimization (FSO) method, including band (1, 4-6, 8) mean, mNDWI, water body frequency, and geometric features (length, width, density, asymmetry, circularity, compactness, shape index, boundary index, and rectangular fitting degree). Then, each block object in the potential water body of the aquaculture pond (the third potential distribution range) is distinguished as a coastal aquaculture pond and a non-aquaculture pond using a nearest neighbor classifier, thereby extracting high-precision spatial distribution data of the coastal aquaculture pond.

[0102] In the embodiment of the present application, a Recursive Classification Algorithm for Multi-scale Coastal zone Aquaculture pond (RCAMCA) is used to extract water bodies at multiple scales, and a rule-based classification algorithm and a nearest neighbor classification algorithm are used to process water body objects in a progressive manner, thereby constructing a feature space that combines spectral features, water body index features, water body frequency features, and geometric features. The present application can expand the monitoring of coastal aquaculture activities to an intercontinental scale, thereby obtaining spatiotemporal dynamic information of large-scale coastal aquaculture activities and analyzing geographic spatial characteristics and hotspot patterns.

[0103] The embodiment of the present application innovates a new method for extracting coastal aquaculture ponds based on the traditional multi-dimensional feature extraction method and the fluctuation characteristics of water body frequency. Compared with the existing published data on coastal aquaculture ponds, the present application effectively solves the technical problem of distinguishing aquaculture ponds from surrounding drainage channels and seawall rivers, and successfully improves the misclassification of rice seedling fields and aquaculture ponds, thereby improving the product accuracy.

[0104] The embodiment of the present application innovatively incorporates seasonal fluctuation information of water bodies to enhance the identification of complex water body dynamic changes, thereby greatly reducing the influence of temporary water facilities on the extraction accuracy of intensive aquaculture ponds, successfully interpreting high-precision spatial distribution data of aquaculture ponds in a long time series, and clarifying the spatial location relationship between mangroves and aquaculture ponds, thereby providing basic data support for effectively analyzing and monitoring the influence of typical human activities in coastal areas on mangrove ecosystems.

[0105] Embodiment two,

[0106] As shown in Figure 6 The embodiment of the present application provides an extraction device for a coastal aquaculture pond, which comprises:

[0107] An image acquisition module 1 is configured to acquire a remote sensing image to be recognized.

[0108] A first range acquisition module 2 is configured to acquire coastline data of a target area, and acquire a first potential distribution range according to the coastline data.

[0109] A second range acquisition module 3 is configured to perform water body and non-water body segmentation on the remote sensing image to be recognized in the first potential distribution range, and acquire a second potential distribution range.

[0110] A historical image acquisition module 4 is configured to acquire a plurality of historical remote sensing images of the second potential distribution range, and perform cloud removal processing.

[0111] A third range acquisition module 5 is configured to calculate water body frequencies of each pixel according to the plurality of historical remote sensing images after the cloud removal processing, and acquire a third potential distribution range according to the water body frequencies.

[0112] A recognized image segmentation module 6 is configured to perform segmentation on the remote sensing image to be recognized in the third potential distribution range by using a multi-scale segmentation algorithm, and acquire a plurality of segmented objects.

[0113] An object classification module 7 is configured to perform recognition on each segmented object by using an object-oriented classification algorithm, and acquire a mariculture pond area.

[0114] Based on the above embodiment, in an optional embodiment of the present application, the first range acquisition module 2 comprises:

[0115] A global coastline data acquisition unit is configured to acquire Open street map global coastline data.

[0116] A buffer zone acquisition unit is configured to extract a region with a first preset distance from both sides of the coastline in the target area according to the Open street map global coastline data, and acquire a buffer zone.

[0117] An elevation model acquisition unit is configured to acquire an SRTM V3 digital elevation model.

[0118] A first range acquisition unit is configured to extract a region with an elevation less than a first preset height in the buffer zone according to the SRTM V3 digital elevation model, and acquire a first potential distribution range.

[0119] Preferably, the first preset distance is 30 km, and the first preset height is 10 m.

[0120] Based on the above embodiment, in an optional embodiment of the present application, the second range acquisition module 3 comprises:

[0121] The first segmentation index calculation unit is configured to calculate a segmentation index of each pixel based on the remote sensing image to be identified in the first potential distribution range, wherein the segmentation index is an index preselected for distinguishing water bodies from non-water bodies.

[0122] The second range acquisition unit is configured to acquire a second potential distribution range of the water body region by performing water body and non-water body segmentation on the remote sensing image to be identified in the first potential distribution range using a global threshold adaptive algorithm.

[0123] Preferably, the step of selecting the segmentation index comprises:

[0124] The remote sensing image of the target region and the standard distribution map of the aquaculture pond are acquired.

[0125] The normalized difference water index, the modified normalized difference water index, and the automatic water extraction index of the remote sensing image of the target region are calculated respectively.

[0126] The remote sensing image of the target region is segmented using a global threshold adaptive algorithm based on the three indexes respectively, and the aquaculture pond distribution map of the target region corresponding to the three indexes is acquired.

[0127] The aquaculture pond distribution map of the target region corresponding to the three indexes and the standard distribution map of the aquaculture pond are compared respectively, and the best segmentation effect among the three indexes is selected as the segmentation index.

[0128] Based on the above-mentioned embodiments, in an optional embodiment of the present application, the historical image acquisition module 4 comprises:

[0129] The historical image acquisition unit is configured to acquire a plurality of historical remote sensing images of the past year in the second potential distribution range, and perform radiation calibration, geometric correction, and atmospheric correction to acquire a plurality of ground reflectance data.

[0130] The cloud removal processing unit is configured to perform cloud removal processing on each ground reflectance data through the Google Earth Engine cloud computing platform.

[0131] Based on the above-mentioned embodiments, in an optional embodiment of the present application, the third range acquisition module 5 comprises:

[0132] The effective observation number acquisition unit is configured to acquire the effective observation number of each pixel based on the plurality of historical remote sensing images after cloud removal processing.

[0133] The water body occurrence frequency acquisition unit is configured to calculate the segmentation index of each ground reflectance data after cloud removal processing, and perform water body and non-water body segmentation using a global threshold adaptive algorithm to acquire the water body occurrence frequency of each pixel.

[0134] The water body frequency layer acquisition unit is configured to acquire a water body frequency layer according to the water body occurrence frequency and the effective observation number.

[0135] The third range acquisition unit is configured to acquire a third potential distribution range according to the water body frequency layer, with the water body occurrence frequency of the pond sample point as a threshold.

[0136] On the basis of the above-mentioned embodiments, in an optional embodiment of the present application, the object classification module 7 comprises:

[0137] The feature space acquisition unit is configured to select a feature space through a feature space optimization method, wherein the feature space comprises the mean values of bands 1, 4, 5, 6 and 8, the modified normalized difference water index, the water body frequency and the set feature.

[0138] The object classification unit is configured to classify each object as a culture pond or a non-culture pond through a nearest neighbor classifier according to the feature space of the segmented object, and acquire a culture pond region.

[0139] Embodiment three,

[0140] The embodiment of the present application provides an extraction device of a coastal culture pond, which comprises a processor, a memory and a computer program stored in the memory. The computer program can be executed by the processor to realize the extraction method of the coastal culture pond as described in any one of the embodiments one.

[0141] Embodiment four,

[0142] The embodiment of the present application provides a computer readable storage medium. The computer readable storage medium comprises a stored computer program, wherein the computer readable storage medium controls the device where the computer readable storage medium is located to execute the extraction method of the coastal culture pond as described in any one of the embodiments one when the computer program runs.

[0143] In several embodiments provided by the embodiments of the present application, it should be understood that the disclosed apparatus and method can also be implemented by other manners. The apparatus and method embodiments described above are only illustrative, for example, the flowcharts and block diagrams in the drawings show the possible implementation architecture, function and operation of the apparatus, method and computer program product according to the embodiments of the present application. In this regard, each block in the flowchart or block diagram can represent a module, a program segment or a part of code, which includes one or more executable instructions for implementing the specified logic function. It should also be noted that in some alternative implementation manners, the functions noted in the blocks can also occur in different order from that noted in the drawings. For example, two consecutive blocks can actually be executed substantially in parallel, and sometimes they can also be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagram and / or flowchart, and the combination of blocks in the block diagram and / or flowchart, can be implemented by a dedicated hardware-based system for executing the specified function or action, or can be implemented by a combination of dedicated hardware and computer instructions.

[0144] In addition, each functional module in the embodiments of the present application can be integrated together to form an independent part, or each module can exist independently, or two or more modules can be integrated to form an independent part.

[0145] If the functions are implemented in the form of software function modules and sold or used as independent products, they can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application essentially or the part of the prior art that makes a contribution or the part of the technical solutions can be embodied in the form of a software product, which is stored in a storage medium and includes a number of instructions for causing a computer device (which can be a personal computer, an electronic device, or a network device, etc.) to execute all or part of the steps of the methods described in the embodiments of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk, and various media that can store program codes. It should be noted that in this document, the terms “include”, “contain” or any other variant thereof are intended to cover non-exclusive inclusion, so that the processes, methods, articles or devices that include a series of elements not only include those elements, but also include other elements not explicitly listed or inherent to such processes, methods, articles or devices. Without more limitations, the element defined by the statement “includes a” does not exclude the presence of another identical element in the process, method, article or device that includes the element.

[0146] The terminology used in the description of the embodiments herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. As used in the description of the embodiments and the appended claims, the singular forms "a", "an" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise.

[0147] It should be understood that the term "and / or" as used herein merely describes associated objects, and can exist in three forms, for example, A and / or B can mean that A exists alone, A and B exist together, or B exists alone. In addition, the character " / " herein generally represents an "or" relationship between the front and rear associated objects.

[0148] Depending on context, the word "if" as used herein can be interpreted to mean "when" or "upon" or "in response to determining" or "in response to detecting." Similarly, the phrase "if it is determined" or "if [a stated condition or event] is detected" can be interpreted to mean "upon determining" or "in response to determining" or "upon detecting [the stated condition or event]" or "in response to detecting [the stated condition or event]."

[0149] The "first\second" mentioned in the embodiments is only to distinguish similar objects, and does not represent a specific order of the objects. It can be understood that the "first\second" can be interchanged in a specific order or sequence as appropriate. It should be understood that the objects distinguished by "first\second" can be interchanged as appropriate, so that the embodiments described herein can be implemented in an order other than those illustrated or described herein.

[0150] The above only describes the preferred embodiments of the present application and is not intended to limit the present application. The present application can have various modifications and changes for those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method of extracting from a coastal zone aquaculture pond, characterized in that, The application relates to an aquaculture pond area recognition method and device. The method comprises the following steps: acquiring a remote sensing image to be recognized; acquiring coastline data of a target area, and acquiring a first potential distribution range according to the coastline data; performing water body and non-water body segmentation on the remote sensing image to be recognized in the first potential distribution range, and acquiring a second potential distribution range; acquiring a plurality of historical remote sensing images of the second potential distribution range, and performing cloud removal processing; calculating water body frequencies of each pixel according to the plurality of historical remote sensing images after cloud removal processing, and acquiring a third potential distribution range according to the water body frequencies; performing segmentation on the remote sensing image to be recognized in the third potential distribution range by using a multi-scale segmentation algorithm, and acquiring a plurality of segmentation objects; and performing recognition on each segmentation object by using an object-oriented classification algorithm, so as to acquire the aquaculture pond area. The method comprises the following steps: acquiring a remote sensing image to be recognized; acquiring coastline data of a target area, and acquiring a first potential distribution range according to the coastline data; performing water body and non-water body segmentation on the remote sensing image to be recognized in the first potential distribution range, and acquiring a second potential distribution range; acquiring a plurality of historical remote sensing images of the second potential distribution range, and performing cloud removal processing; calculating water body frequencies of each pixel according to the plurality of historical remote sensing images after cloud removal processing, and acquiring a third potential distribution range according to the water body frequencies; performing segmentation on the remote sensing image to be recognized in the third potential distribution range by using a multi-scale segmentation algorithm, and acquiring a plurality of segmentation objects; and performing recognition on each segmentation object by using an object-oriented classification algorithm, so as to acquire the aquaculture pond area. The method comprises the following steps: acquiring a remote sensing image to be recognized; acquiring coastline data of a target area, and acquiring a first potential distribution range according to the coastline data; performing water body and non-water body segmentation on the remote sensing image to be recognized in the first potential distribution range, and acquiring a second potential distribution range; acquiring a plurality of historical remote sensing images of the second potential distribution range, and performing cloud removal processing; calculating water body frequencies of each pixel according to the plurality of historical remote sensing images after cloud removal processing, and acquiring a third potential distribution range according to the water body frequencies; performing segmentation on the remote sensing image to be recognized in the third potential distribution range by using a multi-scale segmentation algorithm, and acquiring a plurality of segmentation objects; and performing recognition on each segmentation object by using an object-oriented classification algorithm, so as to acquire the aquaculture pond area. The method comprises the following steps: acquiring a remote sensing image to be recognized; acquiring coastline data of a target area, and acquiring a first potential distribution range according to the coastline data; performing water body and non-water body segmentation on the remote sensing image to be recognized in the first potential distribution range, and acquiring a second potential distribution range; acquiring a plurality of historical remote sensing images of the second potential distribution range, and performing cloud removal processing; calculating water body frequencies of each pixel according to the plurality of historical remote sensing images after cloud removal processing, and acquiring a third potential distribution range according to the water body frequencies; performing segmentation on the remote sensing image to be recognized in the third potential distribution range by using a multi-scale segmentation algorithm, and acquiring a plurality of segmentation objects; and performing recognition on each segmentation object by using an object-oriented classification algorithm, so as to acquire the aquaculture pond area. The method comprises the following steps: acquiring a remote sensing image to be recognized; acquiring coastline data of a target area, and acquiring a first potential distribution range according to the coastline data; performing water body and non-water body segmentation on the remote sensing image to be recognized in the first potential distribution range, and acquiring a second potential distribution range; acquiring a plurality of historical remote sensing images of the second potential distribution range, and performing cloud removal processing; calculating water body frequencies of each pixel according to the plurality of historical remote sensing images after cloud removal processing, and acquiring a third potential distribution range according to the water body frequencies; performing segmentation on the remote sensing image to be recognized in the third potential distribution range by using a multi-scale segmentation algorithm, and acquiring a plurality of segmentation objects; and performing recognition on each segmentation object by using an object-oriented classification algorithm, so as to acquire the aquaculture pond area. The method comprises the following steps: acquiring a remote sensing image to be recognized; acquiring coastline data of a target area, and acquiring a first potential distribution range according to the coastline data; performing water body and non-water body segmentation on the remote sensing image to be recognized in the first potential distribution range, and acquiring a second potential distribution range; acquiring a plurality of historical remote sensing images of the second potential distribution range, and performing cloud removal processing; calculating water body frequencies of each pixel according to the plurality of historical remote sensing images after cloud removal processing, and acquiring a third potential distribution range according to the water body frequencies; performing segmentation on the remote sensing image to be recognized in the third potential distribution range by using a multi-scale segmentation algorithm, and acquiring a plurality of segmentation objects; and performing recognition on each segmentation object by using an object-oriented classification algorithm, so as to acquire the aquaculture pond area. ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ 2. The extraction method of the coastal zone culture pond according to claim 1, characterized by, The coastline data of the target area is acquired, and a first potential distribution range is acquired according to the coastline data, specifically comprising: Acquiring Open street map global coastline data; According to the Open street map global coastline data, the region on both sides of the coastline in the target area within a first preset distance is extracted, and a buffer zone is acquired; Acquiring SRTM V3 digital elevation model; According to the SRTM V3 digital elevation model, the region with an elevation less than a first preset height in the buffer zone is extracted, and the first potential distribution range is acquired; Wherein, The first preset distance is 30km; the first preset height is 10m.

3. The extraction method of the coastal zone culture pond according to claim 1, characterized by, A plurality of historical remote sensing images of the second potential distribution range are acquired, and cloud removal processing is performed, specifically comprising: Acquiring a plurality of historical remote sensing images of the second potential distribution range in the past year, and performing radiation calibration, geometric correction and atmospheric correction to obtain a plurality of ground reflectivity data; Through the Google Earth Engine cloud computing platform, each ground reflectivity data is subjected to cloud removal processing.

4. An extraction device for a coastal zone aquaculture pond, characterized in that, The extraction method of the coastal zone aquaculture pond is suitable for executing the extraction method of the coastal zone aquaculture pond according to any one of claims 1 to 3; the extraction device comprises: An image recognition module for acquiring a to-be-recognized remote sensing image; A first range acquisition module for acquiring coastline data of a target area and acquiring a first potential distribution range according to the coastline data; A second range acquisition module for performing water and non-water body segmentation on the to-be-recognized remote sensing image in the first potential distribution range to acquire a second potential distribution range: A historical image acquisition module for acquiring a plurality of historical remote sensing images of the second potential distribution range and performing cloud removal processing; A third range acquisition module for calculating the water body frequency of each pixel according to the plurality of historical remote sensing images after cloud removal processing, and acquiring a third potential distribution range according to the water body frequency: An image segmentation module for segmenting the to-be-recognized remote sensing image in the third potential distribution range using a multi-scale segmentation algorithm to acquire a plurality of segmented objects; An object classification module for identifying each segmented object using an object-oriented classification algorithm to acquire an aquaculture pond region.

5. An extraction apparatus for a coastal zone aquaculture pond, characterized in that, The computer program can be executed by the processor to implement the extraction method of the coastal zone aquaculture pond according to any one of claims 1 to 3.

6. A computer readable storage medium characterized by, The computer readable storage medium comprises a stored computer program, wherein the computer readable storage medium controls the device where the computer readable storage medium is located to execute the extraction method of the coastal zone aquaculture pond according to any one of claims 1 to 3 when the computer program is running.

Images