An adaptive remote sensing image recognition method and system

Abstract

The application discloses a self-adaptive remote sensing image recognition method and system, comprising the following steps: acquiring a first geographic coordinate corresponding to a target region to be recognized in a remote sensing image; sending the first geographic coordinate to a 3D panorama map to acquire a panorama map coordinate set with the first geographic coordinate as the center; calculating a second geographic coordinate closest to the first geographic coordinate in the panorama map coordinate set; and calculating an azimuth angle according to the first geographic coordinate and the second geographic coordinate. Through the acquisition of the coordinate to be recognized of the remote sensing image, the second geographic coordinate of the 3D panorama map is acquired according to the coordinate, the azimuth angle is calculated, the second height value of the region to be recognized is calculated according to the length of the shadow and the landmark building, the pitch angle is obtained, the image information of the 3D panorama map is acquired through the azimuth angle, the pitch angle and the second geographic coordinate, and self-adaptive recognition is performed. In this way, the relatively high resolution of the 3D panorama map can improve the recognition effect.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of remote sensing image recognition, and particularly relates to a self-adaptive remote sensing image recognition method and system. BACKGROUND

[0002] Remote sensing images are collected by satellites on the ground. Due to the high altitude of the satellite, the resolution of the collected images is low, and only basic types of objects can be distinguished from the images, such as green circles representing plants and squares representing buildings, and further identification of objects, such as specific plants, cannot be achieved. That is, the existing remote sensing images have low resolution, and cannot achieve high-precision identification of remote sensing images. SUMMARY

[0003] Therefore, it is necessary to provide a self-adaptive remote sensing image recognition method and system to solve the problem of low recognition accuracy of remote sensing images.

[0004] To achieve the above purpose, the present application provides a self-adaptive remote sensing image recognition method, comprising the following steps:

[0005] Obtaining a first geographic coordinate corresponding to a target region to be identified in a remote sensing image;

[0006] Sending the first geographic coordinate to a 3D panoramic map to obtain a panoramic map coordinate set centered on the first geographic coordinate;

[0007] Calculating a second geographic coordinate closest to the first geographic coordinate in the panoramic map coordinate set;

[0008] Calculating an azimuth angle according to the first geographic coordinate and the second geographic coordinate;

[0009] Obtaining a closest landmark building in a landmark building database to the first geographic coordinate, obtaining a first height value and a third geographic coordinate of the landmark building;

[0010] Obtaining a first shadow area length at the third geographic coordinate position in the remote sensing image and a second shadow area length at the first geographic coordinate position, and calculating a second height value of the region to be identified according to the first shadow area length, the second shadow area length and the first height value;

[0011] Calculating a distance according to the first geographic coordinate and the second geographic coordinate, and calculating a pitch angle according to the second height value;

[0012] Sending the azimuth angle, the pitch angle and the second geographic coordinate to the 3D panoramic map to obtain first image information in the 3D panoramic map;

[0013] Sending the first image information to an object recognition database to obtain an identification result.

[0014] Further, the method further comprises the following steps:

[0015] After obtaining the first image information, obtain second image information of a target region to be recognized in the remote sensing image;

[0016] Compare the first image information with the second image information in similarity, and when the similarity is greater than or equal to a preset value, send the first image information to an article recognition database to obtain a recognition result;

[0017] When the similarity is less than the preset value, calculate a coordinate in the panoramic map coordinate set that is spaced from the second geographic coordinate by a preset value as a new second geographic coordinate, recalculate the azimuth and the pitch angle according to the new second geographic coordinate, and reobtain the first image information, send the first image information to the article recognition database, and obtain the recognition result.

[0018] Further, the similarity comparison further includes color comparison.

[0019] Further, before obtaining the first geographic coordinate, the method further includes the steps of:

[0020] Obtaining a center geographic coordinate of the current remote sensing image and sending the center geographic coordinate to a 3D panoramic map, and obtaining a panoramic map coordinate set according to the current center geographic coordinate;

[0021] Determining an identifiable region in a preset range according to the panoramic map coordinate set, superimposing the region on the remote sensing image, and displaying the region, and triggering remote sensing image recognition by clicking the region.

[0022] Further, the method further includes the steps of:

[0023] The recognition result includes a plurality of recognized articles and a similarity order according to the second height value and the proximity of the plurality of recognized articles, and the similarity order of the plurality of recognized articles is changed.

[0024] Further, the obtaining of the first image information in the 3D panoramic map includes:

[0025] Taking a screenshot of the 3D panoramic map and saving the screenshot as the first image information.

[0026] Further, the sending of the first geographic coordinate to the 3D panoramic map includes:

[0027] Sending the first geographic coordinate and a preset scale to the 3D panoramic map.

[0028] Further, the recognition result is sent to a search engine for searching, a search result is obtained, and the search result is displayed on the remote sensing image.

[0029] Further, the article recognition database is a plant type database.

[0030] The application further provides a self-adaptive remote sensing image recognition system, comprising a memory and a processor, wherein the memory stores a computer program, and the computer program is executed by the processor to realize the steps of the method according to any one of the embodiments of the application.

[0031] Different from the prior art, the technical scheme above obtains the coordinate to be recognized of the remote sensing image, then obtains the second geographic coordinate of the 3D panorama according to the coordinate and calculates the azimuth angle, then calculates the second height value of the region to be recognized according to the flagpole building and the shadow length, thereby calculating the pitch angle, then obtains the image information of the 3D panorama through the azimuth angle, the pitch angle and the second geographic coordinate and performs self-adaptive recognition. In this way, the relatively high resolution of the 3D panorama can be utilized to realize relatively accurate recognition of the object and improve the recognition effect. BRIEF DESCRIPTION OF DRAWINGS

[0032] Fig. 1 The method flowchart of the application is shown in the figure.

[0033] Fig. 2 The display interface schematic diagram of the system recognition of the application is shown in the figure.

[0034] Fig. 3 The method flowchart of another embodiment of the application is shown in the figure. DETAILED DESCRIPTION

[0035] To explain the technical content, structural features, purposes and effects of the technical scheme in detail, the following will be described in detail in combination with specific embodiments and the accompanying drawings.

[0036] Please refer to Figs. 1 to 3The application provides an adaptive remote sensing image recognition method, comprising the following steps: step S101 acquires a first geographic coordinate corresponding to a target region to be recognized in a remote sensing image; the region to be recognized can be a position clicked by a mouse as the region to be recognized. The geographic coordinate is a latitude and longitude coordinate. Then, step S102 sends the first geographic coordinate to a 3D panorama map to acquire a panorama coordinate set with the first geographic coordinate as the center; the 3D panorama map acquires a plurality of panorama coordinates with panorama images within a certain range around the first geographic coordinate, and the coordinate is a set. Each coordinate corresponds to a 3D panorama map. Then, step S103 calculates a second geographic coordinate closest to the first geographic coordinate in the panorama coordinate set; the distance can be calculated by a latitude and longitude calculation function Haversine formula to find the closest panorama coordinate point. Then, step S104 calculates an azimuth angle according to the first geographic coordinate and the second geographic coordinate; the azimuth angle is an angle of deflection of the first geographic coordinate relative to the second geographic coordinate with the second geographic coordinate as the center, and the angle is an angle of deflection relative to the north direction. The 3D panorama map of the application can be a 3D road panorama map, and in some embodiments, can also be a 3D panorama map in a park.

[0037] After the azimuth angle is calculated, step S105 acquires a closest landmark building in a landmark building database with the first geographic coordinate, acquires a first height value and a third geographic coordinate of the landmark building. It needs to be emphasized that the landmark building is a landmark building in each region. The height is high, obvious and prominent. These landmark buildings contain their positions and heights. These data are public data, such as a well-known international center in Fuzhou with a height of about 274 meters, a longitude of 29.0737XXX° and a latitude of 119.3135XXX°, i.e. the third geographic coordinate. Different cities have similar landmark buildings, which can be stored in a database. Then, step S106 acquires a first shadow region length at the third geographic coordinate position in the remote sensing image and a second shadow region length at the first geographic coordinate position, and calculates a second height value of the region to be recognized according to the first shadow region length, the second shadow region length and the first height value. The shadow length is a shadow region formed by sunlight on the building in the remote sensing image. The shadow region can be obtained in the following way: first, convert the panorama image near the landmark building into an HSV / gray space and separate the brightness information, and then perform noise reduction processing (Gaussian blur). Then, according to a set brightness threshold, the shadow region is segmented. The shadow contour boundary is found and recorded, and then the farthest position of the shadow contour relative to the third geographic coordinate is calculated, i.e. the first shadow region length.

[0038] Similarly, the length of the second shadow area can be obtained at the first geographic coordinate location using the same method. Since the sunlight angles are similar within the same area, the height of the shadow area to be identified can be calculated by comparing the shadow area length of the benchmark building with that of the area to be identified. In some embodiments, to avoid deviations in the shadow area length calculated from a single benchmark building, multiple benchmark buildings can be used for calculation, such as three. The second height values ​​calculated from these three benchmark buildings are then averaged for greater accuracy. Of course, the second height value is subsequently used to calculate the pitch angle, so the actual accuracy requirement is relatively low, and a certain degree of error is permissible.

[0039] Then, in step S107, the distance is calculated based on the first and second geographic coordinates (this distance can also be used to calculate the magnification ratio of the panoramic image; a larger magnification ratio is used for farther distances, and a smaller magnification ratio is used for closer distances), and the pitch angle is calculated using the second altitude value. The direction directly opposite the pitch angle is then the top position of the target area to be identified. This achieves adaptive position acquisition in the altitude direction.

[0040] Then, in step S108, the azimuth angle, pitch angle, and second geographic coordinates are sent to the 3D panoramic image to obtain the first image information from the 3D panoramic image; thus, the first image information is the image information of the top position of the target area to be identified. Finally, in step S109, the first image information is sent to the object recognition database to obtain the recognition result. There are many open databases available for object recognition, such as Baidu's general object and scene recognition API interface, as well as many specific databases, such as plant variety databases and building recognition databases. Based on relatively clear panoramic photos, relatively accurate object recognition can be achieved. If Bougainvillea is detected on the road surface, the API feedback interface information is as follows: { "status": "success","data": {"name": "Bougainvillea","scientific_name": "Bougainvillea spectabilis Willd.","description": "Bougainvillea (Bougainvillea spectabilis Willd.) is an angiosperm of the genus Bougainvillea in the family Nyctaginaceae. It is a vine-like shrub; branches and leaves are densely covered with soft hairs; thorns are axillary and downward-curving. Leaves are elliptical or ovate with a rounded base; inflorescences are axillary or terminal; bracts are elliptic-ovate with a rounded to cordate base, and are dark red or light purplish-red in color; perianth is green and narrowly tubular, with spreading yellow lobes; ovary is stalked; fruit is densely hairy; flowering period is between winter and spring."}}. Then, as... Fig. 2 As shown, the recognized content can be displayed at the mouse position on the system interface.

[0041] The above embodiment obtains the coordinates of the object to be identified from the remote sensing image, then uses these coordinates to obtain the second geographic coordinates of the 3D panoramic image and calculates the azimuth angle. Next, it calculates the second height value of the area to be identified based on the landmark building and shadow length, thereby calculating the pitch angle. Finally, it uses the azimuth angle, pitch angle, and second geographic coordinates to obtain image information from the 3D panoramic image and adaptively performs identification. This utilizes the relatively high resolution of the 3D panoramic image to achieve more accurate object identification, improving the recognition effect. It should be noted that since both the remote sensing image and the 3D panoramic image have acquisition times, it is necessary to use remote sensing images and 3D panoramic images acquired around the same time for identification to improve accuracy.

[0042] In some embodiments, the method further includes the following steps: after acquiring the first image information, acquiring the second image information of the target area to be identified in the remote sensing image; comparing the similarity between the first image information and the second image information; if the similarity is greater than or equal to a preset value, sending the first image information to the object recognition database to obtain the recognition result; if the similarity is less than the preset value, calculating the coordinates in the panoramic image coordinate set that are spaced one preset value (the preset value can be a coordinate number, such as the next coordinate after the interval, or a distance, such as the nearest coordinate after a 5-meter interval) from the second geographic coordinates as the new second geographic coordinates; recalculating the azimuth and elevation angles based on the new second geographic coordinates; and reacquiring the first image information, sending the first image information to the object recognition database to obtain the recognition result. For example, if the second image information of the target area shows green plants at the top, but the first image information shows a white billboard due to possible obstruction by billboards in the panoramic image, then the similarity comparison between the first and second image information shows a large difference, and therefore no recognition is performed; the location is changed before recognition is performed again. This can improve the accuracy of image recognition. Image similarity comparison can be performed using existing interfaces and algorithms.

[0043] In some embodiments, the similarity comparison includes color comparison. Color comparison can be performed by extracting the proportion of each color component. Simple color comparison allows for rapid elimination, resulting in higher comparison efficiency and a simple algorithm.

[0044] Furthermore, such as Fig. 3 As shown, before obtaining the first geographic coordinates, the method includes the following steps: Step S201: Obtain the center geographic coordinates of the current remote sensing image and send them to the 3D panoramic image; obtain the panoramic image coordinate set based on the current center geographic coordinates; Step S402: Determine the identifiable area within a preset range based on the panoramic image coordinate set, overlay the area on the remote sensing image and display the area, and click on the area to trigger remote sensing image recognition. This allows for the pre-setting of areas that support remote sensing image recognition, avoiding triggering unidentifiable areas.

[0045] In some embodiments, the step of identifying the object further comprises: the identification result comprises a plurality of identified objects, and the similarity ranking of the plurality of identified objects is changed according to the proximity of the second height value to the plurality of identified objects. In this way, the objects with higher similarity can be placed in front, and the identification is more accurate.

[0046] Further, the obtaining of the first image information in the 3D panoramic map comprises: taking a screenshot of the 3D panoramic map and saving it as the first image information. In this way, the existing 3D panoramic map can be borrowed, and the deployment is more convenient without using the interface API.

[0047] Further, the sending of the first geographic coordinate to the 3D panoramic map comprises: sending the first geographic coordinate and a preset scale to the 3D panoramic map. The scale control is realized to avoid obtaining too many or too few coordinates.

[0048] Further, the identification result is sent to a search engine for searching, the search result is obtained, and the search result is displayed on the remote sensing image. In this way, the user can see the information related to the identification result on the remote sensing image, and the information acquisition is facilitated.

[0049] The present application also provides an adaptive remote sensing image identification system comprising a memory and a processor, wherein the memory stores a computer program, and the computer program is executed by the processor to realize the steps of the method according to any one of the embodiments of the present application. The identification system of the present application obtains the coordinate to be identified of the remote sensing image, then obtains the second geographic coordinate of the 3D panoramic map according to the coordinate, and calculates the azimuth angle, then calculates the second height value of the region to be identified according to the length of the shadow and the flagpole building, thereby calculating the pitch angle, and then obtaining the image information of the 3D panoramic map according to the azimuth angle, the pitch angle and the second geographic coordinate, and adaptively identifying. In this way, the relatively high resolution of the 3D panoramic map can be used to realize the relatively accurate identification of the object, and the identification effect is improved.

[0050] It should be noted that although the above embodiments have been described in the present text, the patent protection scope of the present application is not limited thereby. Therefore, based on the innovative idea of the present application, the changes and modifications of the embodiments described in the present text, or the equivalent structure or equivalent process transformation made by using the content of the present application specification and drawings, directly or indirectly apply the above technical solutions to other related technical fields, are all included in the patent protection scope of the present application.

Claims

1. An adaptive remote sensing image recognition method, characterized in that, Includes the following steps: Obtain the first geographic coordinates corresponding to the target area to be identified in the remote sensing image. The geographic coordinates are latitude and longitude coordinates. Send the first geographic coordinates to the 3D panoramic image to obtain a set of panoramic image coordinates centered on the first geographic coordinates. The 3D panoramic image is a 3D road panoramic image. Calculate the second geographic coordinate that is closest to the first geographic coordinate in the panoramic image coordinate set; The azimuth is calculated based on the first geographic coordinate and the second geographic coordinate. The azimuth is the angle of deflection of the first geographic coordinate relative to the second geographic coordinate with the second geographic coordinate as the center. This angle is the angle of deflection relative to the north direction. Retrieve the nearest benchmark building to the first geographic coordinate from the benchmark building database, and obtain the first height value and third geographic coordinate of that benchmark building; Obtain the length of the first shadow region and the length of the second shadow region at the third geographic coordinate position in the remote sensing image, and calculate the second height value of the region to be identified based on the length of the first shadow region, the length of the second shadow region, and the first height value. The distance is calculated based on the first and second geographic coordinates, and the pitch angle is calculated by combining the distance with the second altitude value. Send the azimuth angle, elevation angle, and second geographic coordinates to the 3D panoramic image to obtain the first image information in the 3D panoramic image. The first image information is sent to the object recognition database, which is a plant type database, and the recognition result is obtained.

2. The adaptive remote sensing image recognition method according to claim 1, characterized in that, It also includes the following steps: After acquiring the first image information, the second image information of the target area to be identified in the remote sensing image is acquired; The first image information and the second image information are compared for similarity. If the similarity is greater than or equal to a preset value, the first image information is sent to the item recognition database to obtain the recognition result. If the similarity is less than a preset value, the coordinates in the panoramic image coordinate set that are an interval of a preset value from the second geographic coordinates are calculated as the new second geographic coordinates. The azimuth and pitch angles are recalculated based on the new second geographic coordinates, and the first image information is reacquired. The first image information is then sent to the object recognition database to obtain the recognition result.

3. The adaptive remote sensing image recognition method according to claim 2, characterized in that, It also includes the step of: the similarity comparison includes color comparison.

4. The adaptive remote sensing image recognition method according to claim 1, characterized in that, The steps preceding obtaining the first geographic coordinates include: Obtain the center geographic coordinates of the current remote sensing image and send them to the 3D panoramic image; obtain the panoramic image coordinate set based on the current center geographic coordinates. Based on the coordinate set of the panoramic image, the identifiable area is determined within a preset range, and the area is overlaid on the remote sensing image and displayed. Clicking on the area triggers remote sensing image recognition.

5. The adaptive remote sensing image recognition method according to claim 1, characterized in that, It also includes the following steps: The recognition result includes multiple recognized items and is sorted according to similarity. The similarity sort of the multiple recognized items is changed according to the proximity of the second height value to the multiple recognized items.

6. The adaptive remote sensing image recognition method according to claim 1, characterized in that, The acquisition of the first image information in the 3D panoramic image includes: Capture a screenshot of the 3D panoramic image and save it as the first image information.

7. The adaptive remote sensing image recognition method according to claim 1, characterized in that, Sending the first geographic coordinates to the 3D panoramic image includes: Send the first geographic coordinates and the preset scale to the 3D panoramic image.

8. The adaptive remote sensing image recognition method according to claim 1, characterized in that: The identification results are sent to a search engine for retrieval, and the retrieved results are displayed on the remote sensing image.

9. An adaptive remote sensing image recognition system, characterized in that: It includes a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, it implements the steps of the method as described in any one of claims 1 to 8.

Images