Optimal remote sensing image matching method and system for cultivated land monitoring

Abstract

The invention relates to an optimal remote sensing image matching method and system for cultivated land monitoring, and relates to the technical field of remote sensing image processing, and the method comprises the steps: making an image drop image set through a satellite remote sensing image set, carrying out the cloud removal of each image drop image, and obtaining an image range drop image, optimal matching is carried out on the basis of the falling images after cloud removal, an optimal image after cloud removal, sorting and fusion is selected from the falling images in the image range is used as a base image of a target image, vector processing is carried out on the base image of the target image according to obtained vector data, and change image spots with vector elements inconsistent with image feature land classes are obtained; according to the method, a large amount of remote sensing image data can be acquired, analyzed and processed in real time, and the optimal remote sensing image is selected from the remote sensing image data, so that the high resolution and the time phase of the monitoring image are ensured, and the quick identification of the cultivated land change condition is realized.

Description

Technical Field

[0001] The present application relates to the technical field of remote sensing image processing, and in particular to an optimal remote sensing image matching method and system for cultivated land monitoring. Background Art

[0002] As the importance of farmland protection continues to increase, traditional methods of farmland protection can no longer meet the needs of modern agricultural development. In recent years, modern information technologies such as satellite remote sensing have played an important role in the dynamic monitoring and supervision of farmland protection. However, in the actual dynamic monitoring of farmland protection in southern China, there are problems such as low image utilization due to the cloudy and rainy weather in the region and inadequate identification of land type information in the monitoring area.

[0003] To solve the above problems, most existing solutions use cloud detection methods when selecting remote sensing images. Cloud detection methods include threshold method, feature analysis method, homomorphic filtering method and cluster analysis method, but these methods often have certain limitations. Specifically, the existing remote sensing image selection method simply removes the cloud area of ​​the remote sensing image. The cloud detection map obtained is not the true optimal remote sensing image, and the image obtained after removing the cloud area is not clear and cannot reflect the actual situation of the cultivated land, so it cannot achieve rapid identification and accurate management of cultivated land changes. It can be seen that the existing technology cannot select the optimal remote sensing image that meets the actual requirements of cultivated land monitoring from a large number of satellite remote sensing images. Summary of the invention

[0004] The present application provides an optimal remote sensing image matching method and system for cultivated land monitoring. In the dynamic monitoring of cultivated land protection, the optimal satellite remote sensing image is selected according to the monitoring content and the monitoring period, combined with various image conditions, and in accordance with the principle of application, for comprehensive monitoring of existing cultivated land. It can acquire, analyze and process a large amount of remote sensing image data in real time, realize rapid identification and precise management of cultivated land changes, ensure high resolution and timely phase of monitoring images, solve the problem that the existing technology cannot select the optimal remote sensing image that meets the actual monitoring requirements from a large number of satellite remote sensing images, and also solve the problem that the available range of images is limited by cloud cover under cloudy and rainy conditions.

[0005] In the first aspect, the present application provides an optimal remote sensing image matching method for cultivated land monitoring, including:

[0006] Acquire a satellite remote sensing image set that meets preset quality requirements, and create an image drop atlas based on the satellite remote sensing image set;

[0007] Perform cloud detection and removal processing on each image in the image atlas through a preset image processing model to obtain a cloud-removed image range atlas;

[0008] According to the preset image matching and image selection algorithm, each image range falling image in the image range falling image set is optimally matched to obtain a target image base map, wherein the target image base map is the optimal image after cloud removal, sorting and fusion selected from the image range falling images;

[0009] Performing vector processing on the target image base map according to the acquired vector data to obtain a change spot, wherein the change spot is a spot where the vector element in the vector data is inconsistent with the image feature land type;

[0010] The results of the monitoring map of the current status of cultivated land are determined based on the change map.

[0011] Optionally, obtain a set of satellite remote sensing images that meet preset quality requirements, including:

[0012] Acquire a first remote sensing image set and a second remote sensing image set from an image source according to preset resolution requirements and image quality requirements;

[0013] Preprocessing and collectively correcting the first satellite remote sensing image set and the second satellite remote sensing image set respectively to obtain an orthophoto image set that meets preset quality requirements after eliminating geometric distortion, to serve as the satellite remote sensing image set;

[0014] The orthophoto image set includes a first orthophoto image set corresponding to the first satellite remote sensing image set and / or a second orthophoto image set corresponding to the second satellite remote sensing image set.

[0015] Optionally, performing cloud detection and removal processing on each image in the image atlas through a preset image processing model to obtain a cloud-removed image range atlas, including:

[0016] Using a cloud detection model in a preset image processing model to perform cloud detection on the image atlas, to obtain a binary cloud mask atlas;

[0017] Performing cloud region recognition and post-processing according to the cloud mask atlas to obtain a batch cloud detection atlas;

[0018] According to each cloud detection image in the batch cloud detection atlas, vectorization processing is performed on the cloud area extracted in combination with the corresponding image drop map to obtain an image range drop map set after the cloud range is removed, and the image range drop map set contains the image range drop map.

[0019] Optionally, according to a preset image matching and image selection algorithm, each image range drop map in the image range drop map set is optimally matched to obtain a target image base map, including:

[0020] Classify and mark each image range drop image in the image range drop image set to obtain marking data corresponding to the image range drop image, wherein the marking data includes a marking field carrying an assigned value;

[0021] According to the preset image matching rules, each image range map is sorted by binary matching of resolution and timeliness according to the value assigned to the tag field, so as to obtain a valid image range map set with an image usage order;

[0022] Perform mosaic fusion processing according to the effective image range falling map set to obtain a fused remote sensing image falling map;

[0023] According to the fused remote sensing image, the automatic image retrieval algorithm is used to perform scene-by-scene retrieval and matching to obtain the optimal image without cloud coverage as the target image base map;

[0024] The values ​​assigned in the tag field include resolution values ​​and phase values, and the phase values ​​are used to determine the timeliness of image acquisition.

[0025] Optionally, classifying and marking each image range drop image in the image range drop image set to obtain marking data corresponding to the image range drop image includes:

[0026] Traverse each image range map and generate a mark type according to the resolution and time phase of the image range map;

[0027] According to the marking type, the resolution and time phase of the image range map are combined to perform value assignment processing to obtain the marking data corresponding to the image range map.

[0028] Optionally, according to a preset image matching rule, each image range map is sorted by resolution and timeliness by binary matching according to the value assigned to the tag field, to obtain a valid image range map set with an image usage order, including:

[0029] The marked data of each image range is used as the element to be sorted, and a bubble sort method is used to perform binary matching pre-sorting based on the marked field to obtain a pre-sorting result, wherein the pre-sorting result includes a resolution pre-sorting result corresponding to the resolution assignment and a timeliness pre-sorting result corresponding to the phase assignment;

[0030] Based on the resolution priority rule, the resolution of each image range map is divided according to the resolution pre-sorting result to obtain a resolution sorted image set with at least one resolution;

[0031] Based on the timeliness priority principle, the timeliness of each image range map in the resolution sorted image set is divided according to the timeliness pre-sorting result to obtain a valid image range map set with an image usage order;

[0032] The pre-sorting results serve as an effective data basis for data processing and matching processes and are used to form an ordered list for fast search.

[0033] Optionally, performing mosaic fusion processing according to the effective image range falling map set to obtain a fused remote sensing image falling map includes:

[0034] Extracting the first-ranked target image range drop map after removing the cloud range from the valid image range drop map set, wherein the target image range drop map includes a target range vector;

[0035] Determine a target cloud removal range according to the target range vector, and select a fused image range falling atlas of a fused image range vector from the valid image range falling atlas according to the target cloud removal range;

[0036] In a fusion order one by one, the target range vector of the target image range map and the image range vectors of each image range map in the fused image range map set are effectively fused until the target image range map is a complete layer to obtain a fused remote sensing image map.

[0037] Optionally, based on the fused remote sensing image, an automatic image retrieval algorithm is used to perform scene-by-scene retrieval and matching to obtain the optimal image without cloud cover, including:

[0038] Using an automatic image retrieval algorithm, searching for images on a target server using image names and storage paths corresponding to the fused remote sensing image drop, to obtain retrieved images;

[0039] The retrieved image is used as the optimal image selected within the final actual monitoring range.

[0040] Optionally, vector processing is performed on the target image base map according to the acquired vector data to obtain a change patch, including:

[0041] Extracting characteristic land types from the target image base map using a semantic segmentation model, and extracting business vector ranges from the acquired vector data;

[0042] The characteristic land type and the business vector range are superimposed to obtain a change map.

[0043] In the second aspect, the present application provides an optimal remote sensing image matching system for cultivated land monitoring, including:

[0044] A remote sensing image set acquisition module is used to acquire a satellite remote sensing image set that meets preset quality requirements;

[0045] An image atlas making module is used to make an image atlas according to the satellite remote sensing image set;

[0046] A cloud removal module is used to perform cloud detection and removal processing on each image in the image atlas through a preset image processing model to obtain a cloud-removed image range atlas;

[0047] The optimal matching module is used to perform optimal matching on each image range falling image in the image range falling image set according to a preset image matching and image selection algorithm to obtain a target image base map, wherein the target image base map is an optimal image selected from the image range falling images after cloud removal, sorting and fusion;

[0048] A vector processing module is used to perform vector processing on the target image base map according to the acquired vector data to obtain a change spot, wherein the change spot is a spot where the vector element in the vector data is inconsistent with the image feature land type;

[0049] The monitoring module is used to determine the results of the monitoring spots of the current status of cultivated land based on the change spots.

[0050] In summary, the embodiment of the present application uses the acquired satellite remote sensing image set to produce an image atlas set, performs cloud removal processing on each image atlas in the image atlas set to obtain an image range atlas, and then performs optimal matching on the basis of the cloud-removed atlas, and uses the optimal image after cloud removal, sorting, and fusion selected from the image range atlas as the target image base map, and performs vector processing on the target image base map according to the acquired vector data to obtain change patches in which vector elements and image feature land types are inconsistent, and determines the results of the cultivated land status monitoring patches according to the change patches. The embodiment of the present application can acquire, analyze and process a large amount of remote sensing image data in real time, select the optimal remote sensing image therefrom, to ensure the high resolution and timely phase of the monitoring image, realize the rapid identification of cultivated land changes, meet the actual cultivated land monitoring requirements, and solve the problem that the prior art cannot select the optimal remote sensing image that meets the actual cultivated land monitoring requirements from a large number of satellite remote sensing images. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0052] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0053] Figure 1 A schematic diagram of a flow chart of an optimal remote sensing image matching method for farmland monitoring provided in an embodiment of the present application;

[0054] Figure 2 This is a schematic diagram of the steps of an optimal remote sensing image matching method for farmland monitoring provided by an optional embodiment of the present application;

[0055] Figure 3 It is a technical roadmap of an optimal remote sensing image matching method for cultivated land monitoring provided by an optional example of this application;

[0056] Figure 4 It is a schematic diagram of an optimal remote sensing image selection system provided by an optional example of this application;

[0057] Figure 5 is an image classification labeling diagram provided by an optional embodiment of the present application;

[0058] Figure 6 This is a schematic diagram of image fusion processing provided by an optional embodiment of the present application;

[0059] Figure 7 It is a use image sequence list diagram provided by an optional embodiment of the present application;

[0060] Figure 8 A structural block diagram of an optimal remote sensing image matching system for farmland monitoring provided in an embodiment of the present application;

[0061] Fig. 9 It is a structural schematic diagram of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0062] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0063] To facilitate the understanding of the embodiments of the present application, further explanation will be given below in conjunction with the drawings and specific embodiments. The embodiments do not constitute a limitation on the embodiments of the present application.

[0064] Figure 1 A schematic diagram of a flow chart of an optimal remote sensing image matching method for farmland monitoring provided in an embodiment of the present application. Figure 1 As shown, the optimal remote sensing image matching method for cultivated land monitoring provided in the embodiment of the present application may specifically include the following steps:

[0065] Step 110, obtaining a satellite remote sensing image set that meets preset quality requirements, and producing an image drop atlas based on the satellite remote sensing image set.

[0066] In this embodiment, the satellite remote sensing image set includes at least two satellite remote sensing images, and the satellite remote sensing images may include but are not limited to: commercially purchased satellite remote sensing images (such as Beijing-2 and Jilin-1) and public welfare satellite remote sensing images (such as Gaofen-1 and Ziyuan-1), etc. Specifically, this embodiment may select satellite remote sensing images that meet the preset quality requirements. Preferably, the resolution of the satellite remote sensing images may cover 0.5 meters to 2 meters, thereby simultaneously meeting the monitoring needs of large data coverage and high resolution requirements.

[0067] In a specific implementation, this embodiment can obtain satellite remote sensing images that meet preset quality requirements as current image data, and then check whether the color and texture of the current image data are abnormal, so as to produce an image drop map that meets the actual image range and obtain an image drop map set.

[0068] Step 120, performing cloud detection and removal processing on each image in the image atlas through a preset image processing model to obtain a cloud-removed image range atlas.

[0069] In this embodiment, the image range drop map set may include one or more image range drop maps after cloud removal. In this embodiment, a pre-trained deep learning model may be used to perform image cloud detection, drop cloud removal, etc. to obtain the image range drop map after cloud removal.

[0070] In a specific implementation, the deep learning model can be a U-Net model improved by a deep residual network. This embodiment can use the model to extract features from the input image map, and then distinguish cloud areas and non-cloud areas based on the extracted features to achieve cloud detection. Then, the cloud area detected is removed to obtain the image range map after cloud removal. As a result, this embodiment realizes the problem of processing cloud shadows in remote sensing images, and solves the problem of limitations caused by cloud cover in some areas such as the southern region due to cloudy and rainy conditions.

[0071] Step 130, according to the preset image matching and image selection algorithm, optimal matching is performed on each image range falling map in the image range falling map set to obtain the target image base map.

[0072] The target image base map is the optimal image selected from the image range map after cloud removal, sorting and fusion.

[0073] In this embodiment, the optimal image can also be called the best / optimal remote sensing image; the image matching principle can be understood as the image selection principle, mainly including but not limited to resolution priority and timing priority, and the image selection algorithm can be a bubble algorithm, which is not limited in this embodiment.

[0074] In this embodiment, all the falling images excluding the range affected by cloud cover are sorted. The falling images can be sorted in a certain order based on the image selection principle, such as in ascending order, and then the sorted image range falling images are fused, and then sorted again according to a certain algorithm to obtain the best matching image as the target image base map.

[0075] In a specific implementation, this embodiment can first sort the dropped images according to resolution priority, and then perform secondary sorting according to time sequence to obtain a set of dropped images with a certain order, and then use a bubbling algorithm to sort the set of dropped images, thereby providing an orderly data foundation for subsequent data processing and matching processes, and optimizing subsequent matching and retrieval operations through sorting.

[0076] Step 140, performing vector processing on the target image base map according to the acquired vector data to obtain change spots.

[0077] The change spots are spots where the vector elements in the vector data are inconsistent with the image feature land types.

[0078] Step 150, determining the results of the cultivated land status monitoring spots according to the change spots.

[0079] A unified description of steps 140 to 150 is given as follows:

[0080] In this embodiment, the image feature land type may include but is not limited to cultivated land type, etc. The land type of the change map can be set according to actual usage requirements. For example, when applied to cultivated land change monitoring, the cultivated land type can be identified from the target image base map; vector data is mainly used to determine the business vector range, referring to the monitoring range, such as the cultivated land range, etc.

[0081] In the specific implementation, for the monitoring of cultivated land changes, this embodiment uses the cultivated land scope of the latest land change survey results as the monitoring scope, extracts the land type in the optimal image through vector + image interpretation, and then superimposes it with the business vector range, so as to find the spots where the vector elements are inconsistent with the image feature land type, and realize the extraction of change spots where the remote sensing image features are inconsistent with the cultivated land type in the latest annual land survey database.

[0082] After the change spots are extracted in this embodiment, the results of the cultivated land status monitoring spots can be obtained according to the change spots, thereby realizing dynamic and accurate monitoring of changes in the scope of regional cultivated land.

[0083] It can be seen that the embodiment of the present application uses the acquired satellite remote sensing image set to produce an image atlas set, performs cloud removal processing on each image atlas in the image atlas set to obtain an image range atlas, and then performs optimal matching on the basis of the cloud-removed atlas, and uses the optimal image after cloud removal, sorting, and fusion selected from the image range atlas as the target image base map, and performs vector processing on the target image base map according to the acquired vector data to obtain change patches in which the vector elements are inconsistent with the image feature land types, and determines the results of the cultivated land status monitoring patches according to the change patches. The embodiment of the present application can acquire, analyze and process a large amount of remote sensing image data in real time, select the optimal remote sensing image therefrom, to ensure the high resolution and timely phase of the monitoring image, realize the rapid identification of cultivated land changes, meet the actual cultivated land monitoring requirements, and solve the problem that the prior art cannot select the optimal remote sensing image that meets the actual cultivated land monitoring requirements from a large number of satellite remote sensing images.

[0084] Reference Figure 2 , shows a schematic flow chart of the steps of an optimal remote sensing image matching method for farmland monitoring provided by an optional embodiment of the present application. The method may specifically include the following steps:

[0085] Step 210, obtaining a satellite remote sensing image set that meets preset quality requirements, and producing an image drop atlas based on the satellite remote sensing image set.

[0086] In an optional embodiment, the embodiment of the present application obtains a satellite remote sensing image set that meets preset quality requirements, which may specifically include: obtaining a first remote sensing image set and a second remote sensing image set from an image source according to preset resolution requirements and image quality requirements; preprocessing and set correction are performed on the first satellite remote sensing image set and the second satellite remote sensing image set respectively to obtain an orthophoto image set that meets the preset quality requirements after eliminating geometric distortion, as the satellite remote sensing image set; wherein the orthophoto image set includes a first orthophoto image set corresponding to the first satellite remote sensing image set and / or a second orthophoto image set corresponding to the second satellite remote sensing image set.

[0087] In this embodiment, the first remote sensing image set may include commercially purchased satellite remote sensing images obtained through commercially purchased satellites (such as Beijing-2 and Jilin-1); the second remote sensing image set may include public welfare satellite remote sensing images obtained through public welfare satellites (such as Gaofen-1 and Ziyuan-1).

[0088] In a specific implementation, this embodiment processes the first remote sensing image set and the second remote sensing image set, and selects a satellite remote sensing image set therefrom. Each satellite remote sensing image in the satellite remote sensing image set may be composed of optimal images that meet the requirements selected from the two remote sensing image sets. Exemplarily, each optimal image may all come from the same remote sensing image set, or may come from the first remote sensing image set and the second remote sensing image set, such as mainly from the first remote sensing image set, and using the remote sensing images of the first remote sensing image set as a supplement to the areas not covered by the images in the first remote sensing image set. This example does not limit the method of selecting the satellite remote sensing image set from the first remote sensing image set and the second remote sensing image set.

[0089] In the specific implementation, refer to Figure 3 The optimal remote sensing image matching method for cultivated land monitoring implemented in this embodiment mainly involves steps such as data collection and quality inspection, cloud detection and removal based on deep learning models, automatic matching of optimal remote sensing images, automatic selection algorithm of optimal remote sensing images, and extraction of cultivated land change spots based on semantic segmentation models.

[0090] In the relevant technologies, the actual dynamic monitoring of farmland protection in some areas has problems such as low image utilization rate and inadequate identification of land type information in the monitoring area due to the cloudy and rainy weather in the southern region. In addition, some areas also have the phenomenon of monitoring cultivated land maps by manually selecting monitoring images and identifying cultivated land changes scene by scene, which leads to untimely response.

[0091] In order to improve the efficiency and accuracy of farmland protection, this embodiment builds an efficient and intelligent farmland protection dynamic monitoring best / optimal image selection system for the above steps. Figure 4 As shown in the figure, the system mainly includes: data integration and preprocessing module, image declouding module, optimal image automatic matching module, image automatic selection module and change spot extraction module. Among them, the data integration and preprocessing module mainly involves the acquisition and preprocessing of commercial satellite remote sensing images and public welfare satellite remote sensing images, monitoring-related vector range data and thematic management data.

[0092] For data collection and quality inspection, this embodiment can select satellite remote sensing images that meet the resolution requirements and image quality requirements from the acquired satellite remote sensing images. Preferably, the acquired satellite remote sensing images have a resolution of 0.5 meters to 2 meters, thereby meeting the monitoring needs of large amounts of data coverage and high resolution requirements at the same time.

[0093] In actual processing, after acquiring the satellite remote sensing image, this embodiment can perform preliminary processing on the satellite remote sensing image, and the processing process includes but is not limited to: image registration, noise removal, etc. Then, the satellite remote sensing image is geometrically corrected to produce an orthophoto to eliminate geometric distortion in the image and ensure the accuracy of its geographical location and scale.

[0094] In actual implementation, refer to Figure 3 After acquiring the satellite remote sensing image set in this embodiment, each satellite remote sensing image in the satellite remote sensing image set can be used as the current image, and then the color and texture of the current image can be checked to see if they are abnormal, and an image map that conforms to the actual image range can be produced. Among them, the monitoring image quality requirements include but are not limited to: ① The image should be free of large-area noise and stripes, and distorted images should be avoided as much as possible; ② The image color is natural, the texture is clear, and there is no blurring and ghosting; ③ The image should have clear details of the ground objects, moderate contrast, clear layers, and basically balanced colors, and can clearly show the surface coverage and boundaries.

[0095] In an optional manner, this embodiment produces an image drop map set based on the satellite remote sensing image set, which may specifically include: performing color and texture detection processing based on the satellite remote sensing image set to produce an image drop map that conforms to the actual image range.

[0096] In specific implementations, image mapping can be understood as vector mapping. In addition to key information such as image data name, acquisition date, storage path, etc., it can also contain vector information. For example, if a satellite remote sensing image contains the monitoring range of the target area, the corresponding monitoring range can be outlined in the satellite remote sensing image in a vector manner, and relevant information (such as area name, whether there are cloud areas, etc.) can be identified.

[0097] Step 220 , using a cloud detection model in a preset image processing model to perform cloud detection on the image atlas to obtain a binary cloud mask atlas.

[0098] Step 230, performing cloud region recognition and post-processing according to the cloud mask atlas to obtain a batch cloud detection atlas.

[0099] Step 220 and step 230 are described uniformly:

[0100] In a specific implementation, this embodiment can use the U-Net model improved by the deep residual network to build a cloud detection model, and perform a series of model training in advance to obtain a trained cloud detection model. Figure 3 As shown, in this embodiment, each image drop in the image drop set is input into the cloud detection model for cloud detection. The residual module of the model continuously downsamples and extracts the deep features of the input image to improve the detection accuracy of broken clouds and thin clouds. These features can capture the multi-scale information and contextual relationship of the image.

[0101] The encoder in the model is then used to extract features from the input image. Each layer contains a convolutional layer, a batch normalization layer, and a ReLU activation function. During the decoding process, the features extracted by the encoder are recombined and upsampled to the original image size through upsampling and convolution operations to generate a high-resolution segmentation map. The decoder part also uses convolutional layers, batch normalization layers, and ReLU activation functions to ensure effective transmission and fusion of features.

[0102] The decoded feature map is then fused with the input image and convolved again. The model outputs a binary result of whether each pixel is a cloud, and a binary cloud mask is obtained to distinguish between cloud and non-cloud areas. Finally, the output result is post-processed by denoising and boundary smoothing to obtain a batch of cloud detection images.

[0103] Step 240, according to each cloud detection image in the batch cloud detection atlas, vectorization processing is performed on the cloud area extracted in combination with the corresponding image drop map to obtain an image range drop map set after the cloud range is removed.

[0104] Among them, the image range atlas set includes image range atlas.

[0105] In the specific implementation, refer to Figure 3 ,This implementation performs vector processing on the cloud area extracted from each cloud detection map, creates a buffer to solve the problem of small object boundary information and broken small clouds, sets the width parameter of the buffer (i.e., the outward expansion range), removes the area covered by clouds in the image range map, and forms an image range map without cloud coverage after topological inspection.

[0106] Reference Figure 4 As shown in the figure, after the remote sensing image is acquired, the image declouding module performs declouding processing. This module mainly involves cloud detection model to perform cloud masking, vectorize the cloud coverage range, and combine the image range mapping data to form effective image range mapping data, so as to reduce the impact of inaccurate land recognition caused by cloud coverage for subsequent monitoring.

[0107] In actual implementation, the image after cloud detection can be combined with the image drop map of the actual image range, the extracted cloud area can be vectorized, a buffer zone can be created to solve the problems of small object boundary information and broken small clouds, the width parameter of the buffer zone (i.e. the outward expansion range) and the format and path of the output layer can be set, and then the batch cloud detection map can be matched with the corresponding image drop map range, the cloud detection range can be removed from the drop map, and the image drop map with the cloud coverage removed can be formed after topological inspection.

[0108] Step 250, classify and mark each image range falling image in the image range falling image set to obtain marking data corresponding to the image range falling image.

[0109] The tag data includes a tag field carrying an assignment, and the assignment in the tag field includes a resolution assignment and a phase assignment, and the phase assignment is used to determine the timeliness of image mapping.

[0110] In a specific implementation, each image range map has a corresponding resolution and phase (preferably, the phase can be used to monitor the current quarter and determine the timeliness of the image range map). This embodiment sorts the image range maps mainly with reference to their resolution and phase. Therefore, before sorting the image range maps, they can be classified and marked according to the resolution of the image range map, and classification mark fields can be set for different resolutions, and values ​​can be assigned to obtain resolution assignments. Then, on this basis, classification and marking can be further performed according to the phase, and classification mark fields can be set for different phases at the same resolution, and values ​​can be assigned to obtain phase assignments.

[0111] Optionally, this embodiment classifies and marks each image range drop map in the image range drop map set to obtain marking data corresponding to the image range drop map, which may include: traversing each image range drop map and generating a marking type according to the resolution and time phase of the image range drop map; according to the marking type, performing value processing in combination with the resolution and time phase of the image range drop map to obtain the marking data corresponding to the image range drop map.

[0112] For example, refer to Figure 5 The classification mark operation flow chart is shown. The specific process of the optimal image automatic matching module is as follows: ① Classification mark: input image map data, traverse the image resolution and time phase recorded in the image map, divide the image map into 6 types of mark types, and then assign values ​​to the classification mark fields (for example: "1", "2", "3", "4", "5", "6"). 1 corresponds to image resolution (imagegsd, the same below) ≦ 0.5, and the time phase (scenetime, the same below) is the current quarter of monitoring; 2 corresponds to 0.5 < resolution ≦ 0.75, and the time phase is the current quarter of monitoring; 3 corresponds to 0.75 < resolution ≦ 0.8, and the time phase is the current quarter of monitoring; 4 corresponds to resolution ≦ 0.5, and the time phase is the previous quarter of monitoring; 5 corresponds to 0.5 < resolution ≦ 0.8, and the time phase is the previous quarter of monitoring; 6 corresponds to the resolution > 0.8 in the remaining images.

[0113] Step 260 , according to the preset image matching rules, each image range map is sorted by resolution and timeliness according to the value assigned to the tag field, to obtain a valid image range map set with an image usage order.

[0114] In the specific implementation, refer to Figure 3The image matching rule is also called the principle of image selection, which includes two aspects: resolution and timeliness. This embodiment prioritizes resolution and timeliness according to the resolution assignment and phase assignment in each image range map, and performs binary matching sorting on each image range map. For example: images with a resolution of 0.5m are used first, and the timeliness of images is sorted from new to old (taking the monitoring time in June as an example, images of June, May, and April of the quarter are obtained, and they are sorted as June, May, and April); if there is no coverage with a resolution of 0.5m, it is supplemented with images with a resolution better than 1m, and the timeliness of images is sorted from new to old (June, May, April). In this way, a valid image range map set with a usage order is obtained.

[0115] Optionally, this embodiment performs binary matching sorting of resolution and timeliness on each image range map according to the preset image matching rule and the assignment of the tag field to obtain a valid image range map set with an image usage order, which may include: taking the tag data of each image range map as the element to be sorted, and using the bubble sort method to perform binary matching pre-sorting based on the tag field to obtain a pre-sorting result, wherein the pre-sorting result includes a resolution pre-sorting result corresponding to the resolution assignment and a timeliness pre-sorting result corresponding to the phase assignment; based on the resolution priority rule, performing resolution division on each image range map according to the resolution pre-sorting result to obtain a resolution sorted image set of at least one resolution; based on the timeliness priority principle, performing timeliness division on each image range map in the resolution sorted image set according to the timeliness pre-sorting result to obtain a valid image range map set with an image usage order; wherein the pre-sorting result is used as an effective data basis for the data processing and matching process to form an ordered list for quick search.

[0116] In the specific implementation, this embodiment uses the marked data of each image range as the element to be sorted, and then uses the bubble sort method to pre-sort the fields of the map, sorting them in the order of higher resolution and updated phase to form an ordered list, and outputs the sorted image map. Specifically, based on the principle of monitoring image use, the colon sort method is used to automatically match high-resolution optical remote sensing images that meet the monitoring requirements as the monitoring base map. By using the bubble sort method to pre-sort all the fields of the map, an orderly data foundation can be provided for the subsequent data processing and matching process. By optimizing the subsequent matching and retrieval operations through sorting, not only the efficiency of data processing is improved, but also the accuracy of data retrieval and matching is enhanced, providing strong support for the optimization of the entire data processing process.

[0117] For example, refer to Figure 4The system and modules shown, the optimal image automatic matching module mainly involves sorting the image drop based on the principle of using images to generate effective images that are most suitable for generating change spots. Specifically, after obtaining the image drop after cloud removal, firstly classify and mark all the drop images, add the mark fields in turn and assign specific values, and then use the bubble sorting method to pre-sort the fields of the drop images, sort them in the order of higher resolution and updated time phase, form an ordered list, and output the sorted image drop.

[0118] Furthermore, for bubble sorting, this embodiment first regards the classification mark, resolution and phase fields of the image as elements to be sorted, and sorts these fields in the order of larger sequence numbers in different classification marks, newer phases in the same classification mark, and higher resolutions in the same phase, to form an ordered list. When performing image matching, ordered fields can speed up the search, and more efficient search algorithms such as binary search can be used instead of linear search. In addition, sorting can also help to more easily identify and process duplicate or similar image data, thereby improving the efficiency of data cleaning and deduplication.

[0119] Step 270, performing mosaic fusion processing according to the effective image range falling map set to obtain a fused remote sensing image falling map.

[0120] In the specific implementation, this embodiment uses each image in the effective image range image set matched in the previous step as the optimal remote sensing image, and then performs mosaic fusion processing on the optimal remote sensing image. The image image fusion within the monitoring range is performed according to the sorted image data. For example, the image image fusion processing can refer to Figure 6 As shown above, Figure 6 The Y of each image is the part of the image range above without the cloud, and A, B, C, and D are the images after sorting. The list of sorted images can be referred to Figure 7 shown.

[0121] In an optional embodiment, the embodiment of the present application performs mosaic fusion processing according to the effective image range atlas to obtain a fused remote sensing image atlas, which may specifically include: extracting the first-ranked target image range atlas with cloud removal range from the effective image range atlas, the target image range atlas containing a target range vector; determining the target cloud removal range according to the target range vector, and selecting a fused image range atlas of the fused image range vector from the effective image range atlas according to the target cloud removal range; and performing effective range fusion on the target range vector of the target image range atlas and the image range vectors of each image range atlas in the fused image range atlas in a one-by-one fusion order until the target image range atlas is a complete layer to obtain a fused remote sensing image atlas.

[0122] In the related technology, for cloud detection in images, traditional cloud detection methods include threshold method, feature analysis method, homomorphic filtering method and cluster analysis method, but these methods often have certain limitations. In recent years, cloud detection methods based on deep learning have gradually become mainstream. For example, the U-Net method is a convolutional neural network cloud detection method with a U-shaped structure, which has strong feature extraction and reconstruction capabilities. It can accurately locate and segment cloud areas in cloud detection. By inputting the remote sensing image to be detected into the trained U-Net model, the model outputs the binary result of whether each pixel is a cloud, and the output result is post-processed, such as denoising and boundary smoothing, to obtain the final cloud detection map. However, although the existing cloud detection method can segment the cloud area, after segmenting the cloud area, the range of the cloud area occlusion is also segmented, and the occluded part cannot be used. There are obvious vacant areas, which will not reflect the actual status of the cultivated land.

[0123] To solve the above problem, this embodiment uses image fusion to remove the cloud area while completely retaining the blocked area. Specifically, the effective image fusion data within the range after the image fusion processing can be referred to Figure 6 As shown. The specific fusion process includes: removing the cloud range from the first-ranked image range vector, and fusing the image range vectors in order one by one. Exemplarily, when the second-ranked image range vector also has no effective range in the area due to cloud removal, the third-ranked image range vector is fused in succession, and so on, until the first-ranked image range vector in the monitoring area is a complete layer, and a fused remote sensing image range vector is obtained; for example, when there is no gap in the first-ranked image range vector (the image has no clouds), directly proceed to the next step and use it as a fused remote sensing image range vector. Therefore, by performing image range vector fusion, this embodiment obtains an optimal image range vector without clouds and gaps, which is used as a fused remote sensing image range vector for subsequent scene-by-scene retrieval and matching.

[0124] In actual implementation, combined with Figure 6 A detailed supplementary description is given: For the fusion of the image range falling map, this embodiment mainly takes the image range falling map ranked first as the basis and takes it as the target image range falling map, that is, Figure 6 In the range vector represented by the image A before processing, the part of the area without effective range due to cloud removal can be used as the target range vector (refer to Figure 6 The Y area of ​​the image A in the middle is missing due to cloud removal). Then, according to the target range vector, other image range images are selected from the valid image range image set as the target image range image. Specifically, for the target range vector, the image range vectors of other image range images are traversed in the sorting order to determine whether there is no valid range in the area corresponding to the target range vector due to cloud removal. For example, the image range image ( Figure 6 In the image before processing (B), there is no situation where there is no effective range in the area corresponding to the target range vector due to cloud removal, that is, the image range map has an effective range, then the effective range of the image range map is merged with the target range vector to make up for the gap in the target image range map caused by cloud removal. The subsequent image range maps are judged and merged in sequence (refer to Figure 6 The map C and map D in the figure are supplemented) until the missing map data is a complete layer to obtain the fused remote sensing image map.

[0125] The effective range can be understood as: Figure 6 In the figure A and the figure B, the Y areas of the two figures are not completely overlapped or identical. The position of the Y area of ​​figure A in figure B is regarded as the non-overlapping part and is understood as the effective range, which is used to supplement the cloud removal gap in the Y area of ​​figure A.

[0126] Step 280, based on the fused remote sensing image, an automatic image retrieval algorithm is used to perform scene-by-scene retrieval and matching to obtain the optimal image without cloud coverage as the target image base map.

[0127] In the specific implementation, this embodiment uses the image automatic retrieval algorithm to match the image area within the effective image range according to the fused image data, and retrieves the matching image scene by scene in the range order. Specifically, the image automatic retrieval algorithm is used to search for images on the server by image name and storage path, and the retrieved image is used as the image map selected in the final actual monitoring range, and an optimal image base map without cloud coverage is output, that is, the target image base map.

[0128] In an optional embodiment, the embodiment of the present application uses an automatic image retrieval algorithm to perform scene-by-scene retrieval and matching based on the fused remote sensing image map to obtain an optimal image without cloud coverage, which may include: using an automatic image retrieval algorithm to search for images on a target server using the image name and storage path corresponding to the fused remote sensing image map to obtain a retrieved image; and using the retrieved image as the optimal image selected within the final actual monitoring range.

[0129] In the specific implementation, in order to realize the use of automatic image retrieval algorithm, images are searched on the target server. This embodiment first automatically captures the image data of the server, and uses the hash list and string matching (Knuth-Morris-Pratt, KMP) algorithm to perform efficient image matching to achieve rapid retrieval and accurate matching of image resources, and improve the efficiency and accuracy of data processing. The acquired image data includes the image name and storage path, and this information is stored in a hash list. Among them, the selection of the hash list is based on its fast data access and retrieval capabilities, which is crucial for the subsequent matching process.

[0130] In this embodiment, the core matching algorithm KMP algorithm is used for the matching process of the image name, and the invalid backtracking of the pattern string is avoided by constructing a partial matching table (prefix function), thereby improving the matching efficiency. The partial matching table pre-calculates the longest common element length of the prefix and suffix of each position in the pattern string, so that once a mismatch is found during the matching process, the algorithm can use this table to determine which position the pattern string should fall back to, reducing the number of comparisons. The time complexity of the KMP algorithm is O(n+m), where n is the text length and m is the pattern string length, which significantly improves the matching efficiency. In this application scenario, the KMP algorithm matches the list of image names of the dropped image with the server image list, and filters out the used images and their server addresses. This precise screening mechanism not only improves the efficiency of resource management, but also provides a solid foundation for data analysis and processing. Through this refined data management, efficient use and precise control of image resources can be achieved, ensuring the efficiency and accuracy of the data processing flow.

[0131] In actual implementation, refer to Figure 4 The automatic image selection module shown in the figure mainly involves matching and selecting images layer by layer according to the image sorting results using the image automatic retrieval algorithm according to the effective image range to form an optimal monitoring image without cloud coverage. Specifically, in this embodiment, the system can obtain an optimal image base map without cloud coverage through the image automatic retrieval algorithm to form a map for monitoring the current status of cultivated land. Specifically, input the image drop data with the image usage order, read the specified field name (name, scenetime, imagegsd, YXBH) in the drop data attribute table, and the list of sorted image names is as follows: Figure 7 As shown in the figure, the image mosaic fusion is performed in ascending order according to the sorted image number (YXBH) field. Then, the image automatic retrieval algorithm is used to search the image on the server by image name and storage path, and the retrieved image is used as the image map selected in the actual monitoring range, and an optimal image base map without cloud coverage is output.

[0132] Furthermore, in order to improve the visual effect and quality of the image, the present embodiment can perform color uniformity processing on the stitched image to reduce the color difference between different images caused by factors such as shooting conditions and sensor characteristics, and ensure that the color of the entire image is uniform. In addition, by adjusting contrast, reducing noise, sharpening, and atmospheric correction, the clarity and physical authenticity of the image can be further improved.

[0133] Step 290, performing vector processing on the target image base map according to the acquired vector data to obtain change patches, wherein the change patches are patches where the vector elements in the vector data are inconsistent with the image feature land types.

[0134] Optionally, the above-mentioned vector processing of the target image base map according to the acquired vector data to obtain change spots may specifically include: extracting feature land classes from the target image base map using a semantic segmentation model, and extracting business vector ranges from the acquired vector data; and superimposing the feature land classes and the business vector ranges to obtain change spots.

[0135] In the specific implementation, this implementation realizes the extraction of cultivated land change patches based on the semantic segmentation model. Figure 3 and Figure 4 The change patch extraction module is used to extract change patches within the scope of cultivated land, and provide data support for evaluating the scope of regional cultivated land and ensuring the balance of cultivated land occupation and compensation, and the balance of inflow and outflow. This embodiment uses the cultivated land scope of the latest land change survey results as the monitoring scope, and uses the optimal image of the integrated cultivated land status monitoring obtained above. This embodiment interprets the image as "vector + image". Specifically, after extracting the land type of a single-period image using a semantic segmentation model, it is superimposed with the business vector range to find patches where the vector elements are inconsistent with the image feature land type, and realize the extraction of change patches that are inconsistent with the remote sensing image features and the cultivated land type in the latest annual land survey database.

[0136] Step 300, determining the results of the cultivated land status monitoring spots according to the change spots.

[0137] Therefore, the optimal remote sensing image automatic matching method for monitoring the current status of cultivated land proposed in the embodiment of the present application realizes dynamic and accurate monitoring of changes in the scope of regional cultivated land. This embodiment uses the idea of ​​cloud removal and image sorting by deep learning models, which not only solves the problem that optical remote sensing is easily affected by cloudy and rainy weather, resulting in missing land information data in cloud-covered areas, but also uses the image sorting model to realize dynamic monitoring of cultivated land protection with remote sensing images of the latest phase with the best resolution, ensuring the accuracy and timeliness of monitored land types, providing data support and decision-making support for dynamic supervision of cultivated land protection and ensuring food security, and solving the problem that the existing technology cannot select the optimal remote sensing images that meet the actual requirements of cultivated land monitoring from a large number of satellite remote sensing images.

[0138] In summary, the embodiment of the present application produces an image drop atlas based on the acquired satellite remote sensing image set, and then uses the cloud detection model in the preset image processing model to perform cloud detection on the image drop atlas to obtain a binary cloud mask atlas, and then performs cloud area recognition and post-processing according to the cloud mask atlas to obtain a batch cloud detection atlas, and then performs vectorization processing on the cloud area extracted from each cloud detection map in the batch cloud detection atlas in combination with the corresponding image drop map to obtain an image range drop map after the cloud layer range is removed, and then sorts each image range drop map according to resolution and timeliness according to the preset image matching rules. Processing is performed to obtain the label data corresponding to the time-ordered image set and the image range map. According to the label data, the time-ordered image set is binary matched and sorted using a preset bubble sort algorithm to obtain a valid image range map set. Then, mosaic fusion processing is performed based on the valid image range map set to obtain a fused remote sensing image map. On the basis of the fused remote sensing image map, an image automatic retrieval algorithm is used to perform scene-by-scene retrieval and matching to obtain the optimal image without cloud coverage. Vector processing is performed in combination with vector data to obtain change patches that are inconsistent with the image feature land type, thereby determining the patch results of the current status of cultivated land monitoring. The optimal remote sensing image automatic matching method for the current status of cultivated land proposed in this embodiment is intended to solve the problem that the high-quality optical satellite image data available under cloudy and rainy climate conditions in southern China is extremely limited, and the scope of cultivated land cannot be accurately identified in dynamic monitoring of cultivated land protection. This embodiment first uses a deep learning model to detect and remove clouds to deal with cloud shadow problems in remote sensing images; then, all images including the image range covered by clouds are sorted, and the image fusion is performed in ascending order based on the image selection principle. The fused image scene number is read to automatically load the image, thereby realizing the rapid selection of the optimal remote sensing monitoring image from multi-cloud covered, multi-phase, and multi-resolution images, providing technical support for achieving the total balance of cultivated land and cultivated land protection in my country.

[0139] It should be noted that, for the purpose of simple description, the method embodiments are expressed as a series of action combinations, but those skilled in the art should be aware that the embodiments of the present application are not limited to the described order of actions, because according to the embodiments of the present application, certain steps may be performed in other orders or simultaneously.

[0140] like Figure 8 As shown, the embodiment of the present application also provides an optimal remote sensing image matching system 800 for farmland monitoring, including:

[0141] The remote sensing image set acquisition module 810 is used to acquire a satellite remote sensing image set that meets preset quality requirements;

[0142] An image atlas making module 820 is used to make an image atlas according to the satellite remote sensing image set;

[0143] The cloud removal module 830 is used to perform cloud detection and removal processing on each image in the image atlas through a preset image processing model to obtain a cloud-removed image range atlas;

[0144] The optimal matching module 840 is used to perform optimal matching on each image range falling image in the image range falling image set according to a preset image matching and image selection algorithm to obtain a target image base map, wherein the target image base map is an optimal image selected from the image range falling images after cloud removal, sorting and fusion;

[0145] The vector processing module 850 is used to perform vector processing on the target image base map according to the acquired vector data to obtain a change spot, wherein the change spot is a spot where the vector element in the vector data is inconsistent with the image feature land type;

[0146] The monitoring module 860 is used to determine the results of the cultivated land status monitoring spots according to the change spots.

[0147] Optionally, the remote sensing image set acquisition module 810 includes:

[0148] A remote sensing image set acquisition submodule is used to acquire a first remote sensing image set and a second remote sensing image set from an image source according to preset resolution requirements and image quality requirements;

[0149] The remote sensing image set processing submodule is used to preprocess and set correct the first satellite remote sensing image set and the second satellite remote sensing image set respectively, so as to obtain an orthophoto image set that meets the preset quality requirements after eliminating geometric distortion, so as to serve as the satellite remote sensing image set; wherein the orthophoto image set includes a first orthophoto image set corresponding to the first satellite remote sensing image set and / or a second orthophoto image set corresponding to the second satellite remote sensing image set.

[0150] Optionally, the cloud removal module 830 includes:

[0151] A cloud detection submodule, used to perform cloud detection on the image atlas using a cloud detection model in a preset image processing model to obtain a binary cloud mask atlas;

[0152] An identification and post-processing submodule, used for performing cloud region identification and post-processing according to the cloud mask atlas to obtain a batch cloud detection atlas;

[0153] The vectorization processing submodule is used to perform vectorization processing on the cloud area extracted from each cloud detection image in the batch cloud detection atlas in combination with the corresponding image drop map, so as to obtain an image range drop map set after the cloud range is removed, and the image range drop map set includes the image range drop map.

[0154] Optionally, the optimal matching module 840 includes:

[0155] A classification and marking submodule, used for classifying and marking each image range image in the image range image set to obtain marking data corresponding to the image range image, wherein the marking data includes a marking field carrying an assigned value;

[0156] A sorting submodule is used to perform binary matching sorting of resolution and timeliness on each image range map according to the preset image matching rules and the value of the tag field, so as to obtain a valid image range map set with an image usage order;

[0157] A fusion processing submodule is used to perform mosaic fusion processing according to the effective image range falling map set to obtain a fused remote sensing image falling map;

[0158] The retrieval and matching submodule is used to perform scene-by-scene retrieval and matching based on the fused remote sensing image, using the image automatic retrieval algorithm to obtain the optimal image without cloud coverage as the target image base map; wherein the assignment in the tag field includes resolution assignment and phase assignment, and the phase assignment is used to determine the timeliness of the image mapping.

[0159] Optionally, the classification labeling submodule includes:

[0160] A marker type generating unit, used for traversing each image range map, and generating a marker type according to the resolution and time phase of the image range map;

[0161] The assignment unit is used to perform assignment processing according to the mark type in combination with the resolution and time phase of the image range map to obtain the mark data corresponding to the image range map.

[0162] Optional, sorting submodule, including:

[0163] A pre-sorting unit is used to use the marked data of each image range as the elements to be sorted, and use the bubble sorting method to perform binary matching pre-sorting based on the marked field to obtain a pre-sorting result, wherein the pre-sorting result includes a resolution pre-sorting result corresponding to the resolution assignment and a timeliness pre-sorting result corresponding to the time phase assignment;

[0164] A resolution division unit is used to divide the resolution of each image range map according to the resolution pre-sorting result based on the resolution priority rule to obtain a resolution sorted image set of at least one resolution;

[0165] The timeliness division unit is used to divide the timeliness of each image range map in the resolution sorted image set based on the timeliness priority principle and according to the timeliness pre-sorting result, so as to obtain a valid image range map set with the image usage order; wherein the pre-sorting result is used as the effective data basis of the data processing and matching process to form an ordered list for fast search.

[0166] Optionally, the fusion processing submodule includes:

[0167] A fused image range falling atlas acquisition unit is used to determine a target cloud removal range according to the target range vector, and select a fused image range falling atlas of a fused image range vector from the valid image range falling atlas according to the target cloud removal range;

[0168] The fusion unit is used to perform effective range fusion on the target range vector of the target image range map and the image range vector of each image range map in the fused image range map set in a fusion order one by one, until the target image range map is a complete layer, thereby obtaining a fused remote sensing image map.

[0169] Optionally, the retrieval and matching submodule includes:

[0170] A search unit is used to use an automatic image retrieval algorithm to search for images on a target server using image names and storage paths corresponding to the fused remote sensing image drop map to obtain a retrieved image;

[0171] The optimal image selection unit is used to use the retrieved image as the optimal image selected within the final actual monitoring range.

[0172] Optionally, the vector processing module includes:

[0173] A land type extraction submodule, used to extract characteristic land types from the target image base map using a semantic segmentation model, and to extract a business vector range from the acquired vector data;

[0174] The superposition submodule is used to perform superposition according to the characteristic land type and the business vector range to obtain a change patch.

[0175] It should be noted that the optimal remote sensing image matching system for cultivated land monitoring provided in the embodiment of the present application can execute the optimal remote sensing image matching method for cultivated land monitoring provided in any embodiment of the present application, and has the corresponding functions and beneficial effects of the execution method.

[0176] In a specific implementation, the above-mentioned optimal remote sensing image matching system for cultivated land monitoring can be integrated into a device, so that the device can acquire, analyze and process a large amount of remote sensing image data in real time, select the best / optimal remote sensing image through cloud area processing, etc., for comprehensive monitoring of current cultivated land, as an electronic device, to achieve. The electronic device can be composed of two or more physical entities, or it can be composed of one physical entity, such as a personal computer (PC), a computer, a server, etc., and the embodiment of the present application does not make specific restrictions on this.

[0177] like Fig. 9 As shown, an embodiment of the present application provides an electronic device, including a processor 111, a communication interface 112, a memory 113 and a communication bus 114, wherein the processor 111, the communication interface 112, and the memory 113 communicate with each other through the communication bus 114; the memory 113 is used to store computer programs; the processor 111 is used to implement the steps of the optimal remote sensing image matching method for cultivated land monitoring provided by any of the aforementioned method embodiments when executing the program stored in the memory 113. Exemplarily, the steps of the optimal remote sensing image matching method for cultivated land monitoring may include the following steps: obtaining a satellite remote sensing image set that meets preset quality requirements, and producing an image atlas based on the satellite remote sensing image set; performing cloud detection and removal processing on each image in the image atlas through a preset image processing model to obtain a cloud-removed image range atlas; performing optimal matching on each image range atlas in the image range atlas according to preset image matching and image selection algorithms to obtain a target image base map, wherein the target image base map is an optimal image selected from the image range atlas after cloud removal, sorting and fusion; performing vector processing on the target image base map according to the acquired vector data to obtain a change patch, wherein the change patch is a patch where the vector elements in the vector data are inconsistent with the image feature land type; and determining the cultivated land status monitoring patch results according to the change patch.

[0178] An embodiment of the present application also provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the steps of the optimal remote sensing image matching method for cultivated land monitoring provided in any of the aforementioned method embodiments are implemented.

[0179] It should be noted that, in this article, relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0180] The above description is only a specific implementation of the present application, so that those skilled in the art can understand or implement the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest range consistent with the principles and novel features applied for herein.

Claims

1. An optimal remote sensing image matching method for cultivated land monitoring, characterized in that: include: Acquire a satellite remote sensing image set that meets preset quality requirements, and create an image drop atlas based on the satellite remote sensing image set; Perform cloud detection and removal processing on each image in the image atlas through a preset image processing model to obtain a cloud-removed image range atlas; According to the preset image matching and image selection algorithm, each image range falling image in the image range falling image set is optimally matched to obtain a target image base map, wherein the target image base map is the optimal image after cloud removal, sorting and fusion selected from the image range falling images; Performing vector processing on the target image base map according to the acquired vector data to obtain a change spot, wherein the change spot is a spot where the vector element in the vector data is inconsistent with the image feature land type; The results of the monitoring map of the current status of cultivated land are determined based on the change map.

2. The method according to claim 1, characterized in that Obtain satellite remote sensing imagery that meets preset quality requirements, including: Acquire a first remote sensing image set and a second remote sensing image set from an image source according to preset resolution requirements and image quality requirements; Preprocessing and collectively correcting the first satellite remote sensing image set and the second satellite remote sensing image set respectively to obtain an orthophoto image set that meets preset quality requirements after eliminating geometric distortion, to serve as the satellite remote sensing image set; The orthophoto image set includes a first orthophoto image set corresponding to the first satellite remote sensing image set and / or a second orthophoto image set corresponding to the second satellite remote sensing image set.

3. The method according to claim 1, characterized in that Performing cloud detection and removal processing on each image in the image atlas through a preset image processing model to obtain a cloud-removed image range atlas, including: Using a cloud detection model in a preset image processing model to perform cloud detection on the image atlas, to obtain a binary cloud mask atlas; Performing cloud region recognition and post-processing according to the cloud mask atlas to obtain a batch cloud detection atlas; According to each cloud detection image in the batch cloud detection atlas, vectorization processing is performed on the cloud area extracted in combination with the corresponding image drop map to obtain an image range drop map set after the cloud range is removed, and the image range drop map set contains the image range drop map.

4. The method according to claim 1, characterized in that According to the preset image matching and image selection algorithm, each image range map in the image range map set is optimally matched to obtain the target image base map, including: Classify and mark each image range drop image in the image range drop image set to obtain marking data corresponding to the image range drop image, wherein the marking data includes a marking field carrying an assigned value; According to the preset image matching rules, each image range map is sorted by binary matching of resolution and timeliness according to the value assigned to the tag field, so as to obtain a valid image range map set with an image usage order; Perform mosaic fusion processing according to the effective image range falling map set to obtain a fused remote sensing image falling map; According to the fused remote sensing image, the automatic image retrieval algorithm is used to perform scene-by-scene retrieval and matching to obtain the optimal image without cloud coverage as the target image base map; The values ​​assigned in the tag field include resolution values ​​and phase values, and the phase values ​​are used to determine the timeliness of image acquisition.

5. The method according to claim 4, characterized in that Classifying and marking each image range falling image in the image range falling image set to obtain marking data corresponding to the image range falling image, including: Traverse each image range map and generate a mark type according to the resolution and time phase of the image range map; According to the marking type, the resolution and time phase of the image range map are combined to perform value assignment processing to obtain the marking data corresponding to the image range map.

6. The method according to claim 4, characterized in that According to the preset image matching rules, each image range map is sorted by resolution and timeliness according to the value of the tag field, and a valid image range map set with an image usage order is obtained, including: The marked data of each image range is used as the element to be sorted, and a bubble sort method is used to perform binary matching pre-sorting based on the marked field to obtain a pre-sorting result, wherein the pre-sorting result includes a resolution pre-sorting result corresponding to the resolution assignment and a timeliness pre-sorting result corresponding to the phase assignment; Based on the resolution priority rule, the resolution of each image range map is divided according to the resolution pre-sorting result to obtain a resolution sorted image set with at least one resolution; Based on the timeliness priority principle, the timeliness of each image range map in the resolution sorted image set is divided according to the timeliness pre-sorting result to obtain a valid image range map set with an image usage order; The pre-sorting results serve as an effective data basis for data processing and matching processes and are used to form an ordered list for fast search.

7. The method according to claim 4, characterized in that Mosaic fusion processing is performed according to the effective image range falling map set to obtain a fused remote sensing image falling map, including: Extracting the first-ranked target image range drop map after removing the cloud range from the valid image range drop map set, wherein the target image range drop map includes a target range vector; Determine a target cloud removal range according to the target range vector, and select a fused image range falling atlas of a fused image range vector from the valid image range falling atlas according to the target cloud removal range; In a fusion order one by one, the target range vector of the target image range map and the image range vectors of each image range map in the fused image range map set are effectively fused until the target image range map is a complete layer to obtain a fused remote sensing image map.

8. The method according to claim 4, characterized in that According to the fusion remote sensing image, the automatic image retrieval algorithm is used to perform scene-by-scene retrieval and matching to obtain the optimal image without cloud coverage, including: Using an automatic image retrieval algorithm, searching for images on a target server using image names and storage paths corresponding to the fused remote sensing image drop, to obtain retrieved images; The retrieved image is used as the optimal image selected within the final actual monitoring range.

9. The method according to claim 1, characterized in that: According to the acquired vector data, vector processing is performed on the target image base map to obtain a change spot, including: Extracting characteristic land types from the target image base map using a semantic segmentation model, and extracting business vector ranges from the acquired vector data; The characteristic land type and the business vector range are superimposed to obtain a change map.

10. An optimal remote sensing image matching system for cultivated land monitoring, characterized in that: include: A remote sensing image set acquisition module is used to acquire a satellite remote sensing image set that meets preset quality requirements; An image atlas making module is used to make an image atlas according to the satellite remote sensing image set; A cloud removal module is used to perform cloud detection and removal processing on each image in the image atlas through a preset image processing model to obtain a cloud-removed image range atlas; The optimal matching module is used to perform optimal matching on each image range falling image in the image range falling image set according to a preset image matching and image selection algorithm to obtain a target image base map, wherein the target image base map is an optimal image selected from the image range falling images after cloud removal, sorting and fusion; A vector processing module is used to perform vector processing on the target image base map according to the acquired vector data to obtain a change spot, wherein the change spot is a spot where the vector element in the vector data is inconsistent with the image feature land type; The monitoring module is used to determine the results of the monitoring spots of the current status of cultivated land based on the change spots.

Images