Earth surface panoramic image acquisition method and device and electronic equipment

Abstract

The invention discloses an earth surface panoramic image acquisition method and device and electronic equipment, and belongs to the technical field of agricultural monitoring. The method comprises the following steps: acquiring a plurality of earth surface images based on a target pan-tilt camera; the number of the surface images is at least two, and each surface image comprises a plurality of geometric feature points; obtaining a target geometric model of the first earth surface image based on the geometric feature point pair between each first earth surface image and the second earth surface image; the first earth surface image is any image in the plurality of earth surface images, the second earth surface image is an adjacent image of the first earth surface image, and the geometric feature point pairs are matched geometric feature points in the first earth surface image and the second earth surface image; and carrying out panoramic image splicing operation based on all the target geometric models to obtain an earth surface panoramic image. The earth surface panoramic image acquisition method disclosed by the invention is used for solving the technical problem of poor effect of a mode of acquiring an earth surface panoramic image for agricultural monitoring based on texture features of the earth surface image.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of agricultural monitoring, and particularly relates to a method and device for acquiring a ground panoramic image and an electronic device. BACKGROUND

[0002] In the process of agricultural monitoring, the growth status of crops needs to be observed in real time, and a ground panoramic image covering a large area is acquired for subsequent analysis and application. To meet such a demand, multi-view ground images are usually acquired, and the acquired ground images are stitched to obtain a ground panoramic image.

[0003] However, the prior art usually stitches ground images based on the texture features of the ground images. Due to the phenomena of unobvious texture features, large feature errors or even errors, and error accumulation, the method of acquiring a ground panoramic image for agricultural monitoring based on the texture features of ground images has poor effect. SUMMARY

[0004] The present application aims to at least solve one of the technical problems in the prior art. To this end, the present application provides a method and device for acquiring a ground panoramic image and an electronic device to solve the technical problem that the method of acquiring a ground panoramic image for agricultural monitoring based on the texture features of ground images has poor effect.

[0005] In a first aspect, the present application provides a method for acquiring a ground panoramic image, comprising:

[0006] acquiring a plurality of ground images based on a target pan-tilt camera; the number of ground images is at least 2, and each ground image includes a plurality of geometric feature points;

[0007] acquiring a target geometric model of a first ground image based on a geometric feature point pair between the first ground image and a second ground image; the first ground image is any image in the plurality of ground images, the second ground image is an adjacent image of the first ground image, and the geometric feature point pair is a matching geometric feature point in the first ground image and the second ground image;

[0008] performing a panoramic image stitching operation based on all target geometric models to obtain a ground panoramic image.

[0009] According to the ground panoramic image acquisition method, a plurality of ground images are acquired based on the target holder camera, a target geometric model of each ground image is acquired based on a geometric feature point pair between each ground image and an adjacent ground image, and panoramic image stitching is performed based on all target geometric models to acquire the ground panoramic image. By using the geometric feature point pair between each two adjacent ground images to generate the target geometric model of each ground image, the ground panoramic image is acquired, and compared with the panoramic image acquisition method based on texture features, the effect of acquiring the ground panoramic image for agricultural monitoring is significantly improved.

[0010] According to an embodiment of the present application, the target geometric model of the first ground image is acquired based on the geometric feature point pair between the first ground image and the second ground image, comprising:

[0011] The geometric feature points in the first ground image and the geometric feature points in the second ground image are matched to acquire the geometric feature point pair;

[0012] The first geometric model of the first ground image is acquired based on the number of geometric feature point pairs;

[0013] The target geometric model is acquired by optimizing the first geometric model through the random sample consensus algorithm.

[0014] According to an embodiment of the present application, the first geometric model of the first ground image is acquired based on the number of geometric feature point pairs, comprising:

[0015] In a case where the number of geometric feature point pairs is greater than or equal to the feature point number threshold, the first geometric model of the first ground image is acquired based on the homogeneous coordinates of the geometric feature point pairs.

[0016] According to an embodiment of the present application, the first geometric model of the first ground image is acquired based on the number of geometric feature point pairs, comprising:

[0017] In a case where the number of geometric feature point pairs is less than the feature point number threshold, the first geometric model of the first ground image is acquired based on the ground image resolution, the target coordinates of the first ground image and the target coordinates of the second ground image.

[0018] According to an embodiment of the present application, the target geometric model is acquired by optimizing the first geometric model through the random sample consensus algorithm, comprising:

[0019] The first geometric model is projected to the first panoramic image coordinate system to acquire the second geometric model of the first ground image;

[0020] The distance error of all second geometric models is acquired;

[0021] By using the random sample consensus algorithm, error feature points in the geometric feature points of each first ground surface image are removed based on distance errors, and the second geometric model is recalculated based on the geometric feature points after removing the error feature points;

[0022] In a case where the distance error is less than or equal to the error threshold, the second geometric model corresponding to the distance error is determined as the target geometric model.

[0023] According to an embodiment of the present application, panoramic image stitching is performed based on all target geometric models to obtain a ground surface panoramic image, including:

[0024] Projecting all target geometric models in the first panoramic image coordinate system to the second panoramic image coordinate system to obtain a to-be-filled panoramic image; the to-be-filled image includes a first pixel coordinate set of each first ground surface image, and the first pixel coordinate set includes a plurality of first pixel coordinate points, each first pixel coordinate point being a coordinate of a pixel point of the first ground surface image in the second panoramic image coordinate system;

[0025] For each first pixel coordinate set, a target color value of the first ground surface image is obtained based on the target geometric model;

[0026] Based on all target color values, color filling is performed on the to-be-filled panoramic image to obtain a ground surface panoramic image.

[0027] According to an embodiment of the present application, all target geometric models in the first panoramic image coordinate system are projected to the second panoramic image coordinate system to obtain a to-be-filled panoramic image, including:

[0028] Projecting all target geometric models in the first panoramic image coordinate system to the second panoramic image coordinate system to obtain a first pixel coordinate point of all pixel points of each first ground surface image in the second panoramic image coordinate system;

[0029] Determining a difference between a maximum value and a minimum value of the abscissa of the first pixel coordinate point in all first ground surface images as a width of the to-be-filled panoramic image, and determining a difference between a maximum value and a minimum value of the ordinate of the first pixel coordinate point in all first ground surface images as a height of the to-be-filled panoramic image;

[0030] Based on the height and the width of the to-be-filled panoramic image, all first ground surface images in the second panoramic image coordinate system are framed to obtain the to-be-filled panoramic image.

[0031] According to an embodiment of the present application, for each first pixel coordinate set, a target color value of the first ground surface image is obtained based on the target geometric model, including:

[0032] Based on the target geometric model and the second pixel coordinate point, a plurality of color values are obtained; the second pixel coordinate point is a coordinate of a pixel corresponding to the first pixel coordinate point in the first panoramic image coordinate system;

[0033] The color value in the target coordinate range is determined as the target color value of the first ground surface image.

[0034] In a second aspect, the present application provides a ground surface panoramic image acquisition device, comprising:

[0035] The first acquisition module is configured to acquire a plurality of ground surface images based on a target pan-tilt camera; the number of ground surface images is at least 2, and each ground surface image comprises a plurality of geometric feature points;

[0036] The second acquisition module is configured to acquire a target geometric model of a first ground surface image based on a geometric feature point pair between the first ground surface image and a second ground surface image; the first ground surface image is any image in the plurality of ground surface images, the second ground surface image is an adjacent image of the first ground surface image, and the geometric feature point pair is a matching geometric feature point pair in the first ground surface image and the second ground surface image;

[0037] The third acquisition module is configured to perform panoramic image stitching operation based on all target geometric models to obtain a ground surface panoramic image.

[0038] According to the ground surface panoramic image acquisition device of the present application, a plurality of ground surface images are acquired based on a target pan-tilt camera, a target geometric model of each ground surface image is generated based on a geometric feature point pair between the ground surface image and its adjacent ground surface image, and a ground surface panoramic image is acquired by performing panoramic image stitching operation based on all target geometric models, so that the target geometric model of each ground surface image is generated based on the geometric feature point pair between each two adjacent ground surface images with high geometric stitching accuracy, and then the ground surface panoramic image is acquired by stitching, which significantly improves the effect of acquiring the ground surface panoramic image for agricultural monitoring compared with the method of acquiring the panoramic image based on texture features.

[0039] In a third aspect, the present application provides an electronic device comprising a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to realize the steps of the ground surface panoramic image acquisition method of the first aspect.

[0040] Additional aspects and advantages of the present application will be in part apparent and in part pointed out hereinafter in the description of the application. BRIEF DESCRIPTION OF DRAWINGS

[0041] The above and / or additional aspects and advantages of the present application will become apparent and be readily appreciated from the description of the embodiments, taken in conjunction with the following drawings in which:

[0042] Figure 1 is one of flow diagrams of the ground panoramic image acquisition method provided by the embodiments of the present application;

[0043] Figure 2 is a process diagram of acquiring a plurality of ground images provided by the embodiments of the present application;

[0044] Figure 3 is a second panoramic image coordinate system diagram provided by the embodiments of the present application;

[0045] Figure 4 is a to-be-filled panoramic image diagram provided by the embodiments of the present application;

[0046] Figure 5 is a second flow diagram of the ground panoramic image acquisition method provided by the embodiments of the present application;

[0047] Figure 6 is a structure diagram of the ground panoramic image acquisition device provided by the embodiments of the present application;

[0048] Figure 7 is a structure diagram of the electronic device provided by the embodiments of the present application. DETAILED DESCRIPTION

[0049] The technical solutions in the embodiments of the present application will be clearly described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, but not all of them. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art belong to the scope of protection of the present application.

[0050] The terms "first", "second", and the like in the specification and claims of the present application are used to distinguish similar objects, and are not used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances, so that the embodiments of the present application can be implemented in an order other than those illustrated or described herein, and the objects distinguished by "first", "second", etc. are usually a class, and are not limited to the number of objects, for example, the first object can be one or more. In addition, "and / or" in the specification and claims indicates at least one of the connected objects, and the character " / ", generally indicates that the front and rear associated objects are in an "or" relationship.

[0051] The ground panoramic image acquisition method, device and electronic device provided by the embodiments of the present application will be described in detail below with reference to the drawings, through specific embodiments and their application scenarios.

[0052] The ground panoramic image acquisition method can be applied to a terminal, and can be specifically executed by hardware or software in the terminal.

[0053] The terminal includes, but is not limited to, a portable communication device such as a mobile phone or tablet computer having a touch-sensitive surface (e.g., a touchscreen display and / or touchpad). It will also be appreciated that, in some embodiments, the terminal can not be a portable communication device, but rather a desktop computer having a touch-sensitive surface (e.g., a touchscreen display and / or touchpad).

[0054] In the various embodiments described below, a terminal is described that includes a display and a touch-sensitive surface. However, it will be appreciated that the terminal can include one or more other physical user interface devices such as physical keyboards, mice, and joysticks.

[0055] The ground panoramic image acquisition method provided by the embodiments of the present application can be executed by an electronic device or a functional module or functional entity capable of implementing the ground panoramic image acquisition method in the electronic device. The electronic device mentioned in the embodiments of the present application includes but is not limited to a mobile phone, a tablet computer, a computer, a camera, a wearable device, and the like. The ground panoramic image acquisition method provided by the embodiments of the present application will be described below by taking an electronic device as an execution subject.

[0056] As shown in Figure 1 The ground panoramic image acquisition method includes steps 110, 120, and 130.

[0057] Step 110: Obtain a plurality of ground images based on a target pan-tilt camera; the number of ground images is at least 2, and each ground image includes a plurality of geometric feature points.

[0058] In some embodiments, for the ground surface that needs to be monitored for agriculture, a data acquisition device can be installed on the top of a high tower on the ground surface. The data acquisition device includes a high-definition camera combined with a pan-tilt (target pan-tilt camera), a GPS positioning and measuring device, and the like. The high-definition camera can be controlled by the pan-tilt to rotate around the high tower to take pictures of the ground surface, and a plurality of ordered ground images are collected, as shown in Figure 2 Each ground image has a certain degree of overlap with an adjacent image.

[0059] In some embodiments, the target pan-tilt camera can be controlled to rotate horizontally by 60 degrees to obtain six ground images to be spliced.

[0060] In some embodiments, after a plurality of ground images are obtained, the plurality of ground images can be preprocessed.

[0061] In some embodiments, after a plurality of ground images are obtained, in the case that the ground resolutions of the plurality of ground images are different, a bilinear interpolation or a cubic convolution interpolation method can be used to resample the plurality of ground images to obtain a plurality of ground images with the same ground resolution.

[0062] In some embodiments, after obtaining the plurality of surface images, geometric feature points of each surface image can be obtained.

[0063] In some embodiments, the geometric feature points of each surface image can be obtained based on a Scale Invariant Feature Transform (SIFT) algorithm, a Harris corner point algorithm, or any theoretically feasible algorithm.

[0064] In some embodiments, after obtaining the geometric feature points of each surface image, nearest neighbor point matching can be performed on the geometric feature points on two adjacent surface images to obtain geometric feature point pairs of the two adjacent surface images.

[0065] In actual implementation, the geometric feature point pairs can be matching geometric feature point pairs in the overlapping region of the two adjacent surface images.

[0066] Step 120: obtaining a target geometric model of each first surface image based on the geometric feature point pairs between the first surface image and a second surface image; the first surface image is any one of the plurality of surface images, the second surface image is an adjacent surface image of the first surface image, and the geometric feature point pairs are matching geometric feature points in the first surface image and the second surface image.

[0067] In some embodiments, each of the plurality of surface images can be sequentially determined as the first surface image, and a surface image adjacent to the first surface image can be determined as the second surface image. The second surface image can be a surface image adjacent to the left of the first surface image, or the second surface image can be a surface image adjacent to the right of the first surface image.

[0068] In some embodiments, an initial geometric model of each first surface image can be obtained based on the geometric feature point pairs between the first surface image and the second surface image, and the initial geometric model can be iteratively optimized by a random sample consensus algorithm to obtain the target geometric model.

[0069] Step 130: performing a panoramic image stitching operation based on all the target geometric models to obtain a surface panoramic image.

[0070] In some embodiments, the panoramic image stitching operation can be performed on all the first surface images based on the target geometric model corresponding to each first surface image in all the first surface images to obtain the surface panoramic image.

[0071] In some embodiments, the to-be-filled image can be constructed based on the target geometric model corresponding to each first ground surface image in all the first ground surface images, and the to-be-filled image can be a grayscale image. The filling color value of each pixel point in the to-be-filled image is obtained based on the target geometric model, and the color filling is performed on the to-be-filled panoramic image based on the filling color value, so as to obtain the ground surface panoramic image.

[0072] According to the ground surface panoramic image acquisition method, a plurality of ground surface images are obtained based on the target holder camera, the target geometric model of each ground surface image is obtained based on the geometric feature point pair between the ground surface image and its adjacent ground surface image, and the panoramic image stitching operation is performed based on all the target geometric models to obtain the ground surface panoramic image. The target geometric model of each ground surface image is generated based on the geometric feature point pair with high geometric stitching precision between each two adjacent ground surface images, and then the ground surface panoramic image is obtained by stitching. Compared with the method of acquiring the panoramic image based on the texture feature, the effect of acquiring the ground surface panoramic image for agricultural monitoring is significantly improved.

[0073] In some embodiments, the geometric feature points in the first ground surface image and the geometric feature points in the second ground surface image are matched to obtain the geometric feature point pair, the first geometric model of the first ground surface image is obtained based on the number of geometric feature point pairs, and the target geometric model is obtained by performing optimization processing on the first geometric model through the random sample consensus algorithm.

[0074] In some embodiments, the generation manner of the first geometric model of the first ground surface image can be determined based on the number of geometric feature point pairs between the first ground surface image and the second ground surface image, and the first geometric model is generated based on the generation manner.

[0075] In actual implementation, the first geometric model can be an initial model to be optimized.

[0076] In actual implementation, the basic idea of the random sample consensus (RANSAC) algorithm is as follows: a minimum sample set is randomly extracted from a sample set to initialize a model, the error of the model and the remaining samples is calculated, a sample set with an error less than a threshold value is found, which is called a consistent set. If the number of samples in the consistent set is greater than a preset value, it is considered that the model parameters are correct, and the model is recalculated using the consistent set. The above process is repeated until a preset iteration number is reached or the best model is found.

[0077] In some embodiments, after the first geometric model of the first ground surface image is acquired, some geometric feature point pairs with large errors can be removed by a random sample consensus algorithm, and the first geometric model is continuously acquired based on the remaining geometric feature point pairs, and the error value is calculated. In the case where the error value is greater than the expected target error, the process of removing geometric feature point pairs with large errors and generating the first geometric model is continuously repeated until the error value is less than or equal to the expected target error, and the last acquired first geometric model is determined as the target geometric model.

[0078] According to the ground surface panoramic image acquisition method, a plurality of ground surface images are acquired based on the target pan-tilt camera, and the first geometric model of the first ground surface image is acquired based on the number of geometric feature point pairs between each ground surface image and its adjacent ground surface image. The target geometric model is acquired by optimizing the first geometric model through the random sample consensus algorithm, and the panoramic image stitching operation is performed based on all target geometric models to acquire the ground surface panoramic image. Through the robustness of the random sample consensus (RANSAC) algorithm, a suitable target geometric model can be calculated, and errors caused by incorrect geometric feature point pairs are avoided, thereby improving the accuracy of acquiring the ground surface panoramic image and significantly improving the effect of acquiring the ground surface panoramic image for agricultural monitoring.

[0079] In some embodiments, in the case where the number of geometric feature point pairs is greater than or equal to the feature point number threshold, the first geometric model of the first ground surface image is acquired based on the homogeneous coordinates of the geometric feature point pairs.

[0080] In actual implementation, the feature point number threshold can be a set value. For example, the feature point number threshold can be set based on the number of geometric feature point pairs.

[0081] In some embodiments, in the case where the number of geometric feature point pairs is greater than or equal to the feature point number threshold, the first geometric model of the first ground surface image can be acquired based on the homogeneous coordinates of the geometric feature point pairs. For example, in the case where the feature point number threshold is 10 and the number of geometric feature point pairs is 12, the first geometric model of the first ground surface image can be acquired based on the homogeneous coordinates of the geometric feature point pairs.

[0082] In some embodiments, in the case where the number of geometric feature point pairs is greater than or equal to the feature point number threshold, the first geometric model of the first ground surface image can be acquired based on the following formula:

[0083]

[0084] Wherein, H ia ​​​​​​​​a first geometric model representing the i-th first ground surface image, a geometric transformation between the geometric feature points of the first ground surface image and the geometric feature points of the second ground surface image is represented as x i1 = H ia x i2 , x i1 is the homogeneous coordinate of the geometric feature point of the i-th first ground surface image, x i2 is the homogeneous coordinate of the geometric feature point of the second ground surface image corresponding to the i-th first ground surface image.

[0085] According to the ground surface panoramic image acquisition method, a plurality of ground surface images are acquired based on the target holder camera, in the case that the number of geometric feature point pairs is greater than or equal to the feature point number threshold, the first geometric model of the first ground surface image is acquired based on the homogeneous coordinates of the geometric feature point pairs, the target geometric model is acquired by optimizing the first geometric model through the random sample consensus algorithm, and the panoramic image stitching operation is performed based on all target geometric models to acquire the ground surface panoramic image. Through the robustness of the random sample consensus (RANSAC) algorithm, it is ensured that the appropriate target geometric model can be calculated, and the error model caused by the error geometric feature point pair is avoided, so as to improve the accuracy of acquiring the ground surface panoramic image, and significantly improve the effect of acquiring the ground surface panoramic image for agricultural monitoring.

[0086] In some embodiments, in the case that the number of geometric feature point pairs is less than the feature point number threshold, the first geometric model of the first ground surface image is acquired based on the ground surface image resolution, the target coordinates of the first ground surface image and the target coordinates of the second ground surface image.

[0087] In some embodiments, in the case that the number of geometric feature point pairs is less than the feature point number threshold, the first geometric model of the first ground surface image can be acquired based on the ground surface image resolution, the target coordinates of the first ground surface image and the target coordinates of the second ground surface image. For example, in the case that the feature point number threshold is 10, the number of geometric feature point pairs is 8, the first geometric model of the first ground surface image can be acquired based on the ground surface image resolution, the target coordinates of the first ground surface image and the target coordinates of the second ground surface image.

[0088] In actual implementation, the target coordinates can be the starting ground coordinates of the first ground surface image or the second ground surface image, such as the coordinates of the center pixel point of the first ground surface image or the second ground surface image, or the coordinates of the top-left pixel point of the first ground surface image or the second ground surface image, or the coordinates of the pixel point at any position.

[0089] In some embodiments, in the case that the number of geometric feature point pairs is less than the feature point number threshold, the first geometric model of the first ground surface image can be acquired based on the following formula:

[0090]

[0091] t x =(X j -X i ) / GSD;

[0092] t y =(Y j -Y i ) / GSD;

[0093] wherein H ia represents the first geometric model of the i-th first ground image, (X i , Y i ) represents the target coordinates of the first ground image, (X j , Y j ) represents the target coordinates of the second ground image, and GSD is the ground resolution.

[0094] According to the ground panoramic image acquisition method, a plurality of ground images are acquired based on the target pan-tilt camera, in the case that the number of geometric feature point pairs is less than the threshold of the number of feature points, the first geometric model of the first ground image is acquired based on the ground image resolution, the target coordinates of the first ground image and the target coordinates of the second ground image, the target geometric model is acquired by optimizing the first geometric model through the random sample consensus algorithm, and the panoramic image stitching operation is performed based on all the target geometric models to acquire the ground panoramic image, so as to avoid the stitching failure caused by too few geometric feature points or mis-matching of the geometric feature points, and the effect of acquiring the ground panoramic image for agricultural monitoring is significantly improved.

[0095] In some embodiments, the first geometric model can be projected to the first panoramic image coordinate system to acquire the second geometric model of the first ground image, the distance error of all the second geometric models is acquired, the error feature points in the geometric feature points of each first ground image are removed based on the distance error through the random sample consensus algorithm, and the second geometric model is recalculated based on the geometric feature points after removing the error feature points, and in the case that the distance error is less than or equal to the error threshold, the second geometric model corresponding to the distance error is determined as the target geometric model.

[0096] In actual implementation, the first panoramic image coordinate system can be a coordinate system established with the target coordinates of any one first ground image as the coordinate system origin.

[0097] In some embodiments, the second geometric model of the first ground image can be acquired based on the following formula:

[0098] H1=I;

[0099] Hi = H i-1 H ia ;

[0100] wherein I represents a unit matrix, H1 represents the second geometric model of the first first ground surface image, H i represents the second geometric model of the i-th first ground surface image, H ia represents the first geometric model of the i-th first ground surface image.

[0101] In some embodiments, the distance error of all the second geometric models can be obtained based on the following formula:

[0102]

[0103] wherein G represents the distance error of all the second geometric models, H i represents the second geometric model of the i-th first ground surface image in the N first ground surface images, x i represents the horizontal coordinate of the target coordinate of the i-th first ground surface image, H i+1 represents the second geometric model of the second ground surface image corresponding to the i-th first ground surface image, x i+1 represents the horizontal coordinate of the target coordinate of the second ground surface image corresponding to the i-th first ground surface image.

[0104] In some embodiments, a nonlinear iterative optimization process (Bundle Adjust) can be performed based on the distance error to remove the error feature points in the geometric feature points of each first ground surface image, and the second geometric model is recalculated based on the geometric feature points after removing the error feature points, and the distance error is obtained based on the recalculated second geometric model, and in the case that the distance error is greater than the error threshold, the above process is repeated, and in the case that the distance error is less than or equal to the error threshold, the second geometric model corresponding to the distance error is determined as the target geometric model.

[0105] According to the ground panoramic image acquisition method provided in the embodiments of the present application, a plurality of ground images are acquired based on a target holder camera, a first geometric model of a first ground image is acquired based on the number of geometric feature point pairs, the first geometric model is projected to a first panoramic image coordinate system, and a second geometric model of the first ground image is acquired; distance errors of all second geometric models are acquired; error feature points in the geometric feature points of each first ground image are removed based on the distance errors through a random sample consensus algorithm, and the second geometric model is recalculated based on the geometric feature points after the error feature points are removed; in the case that the distance error is less than or equal to an error threshold, the second geometric model corresponding to the distance error is determined as a target geometric model, and panoramic image stitching operations are performed based on all target geometric models to acquire a ground panoramic image, so that the first geometric model between the first ground image and the second ground image is converted into the second geometric model in the first panoramic image coordinate system, and a global optimization process (Bundle Adjust) is performed to correct the result deviation and closed-loop misplacement caused by the cumulative error of two-by-two stitching, thereby ensuring the stitching geometric precision between every two adjacent ground images and significantly improving the effect of acquiring the ground panoramic image for agricultural monitoring.

[0106] In some embodiments, all target geometric models of the first panoramic image coordinate system can be projected to a second panoramic image coordinate system to acquire a panoramic image to be filled; the panoramic image to be filled includes a first pixel coordinate set of each first ground image, and the first pixel coordinate set includes a plurality of first pixel coordinate points, each first pixel coordinate point being a coordinate of a pixel point of the first ground image in the second panoramic image coordinate system; for each first pixel coordinate set, a target color value of the first ground image is acquired based on the target geometric model; and the panoramic image to be filled is color-filled based on all target color values to acquire the ground panoramic image.

[0107] In actual implementation, as shown in Figure 3 the horizontal coordinate minimum value of the pixel point coordinate of the first ground image corresponding to all target geometric models of the first panoramic image coordinate system can be taken as the vertical axis of the second panoramic image coordinate system, and the horizontal axis of the first panoramic image coordinate system can be taken as the horizontal axis of the second panoramic image coordinate system.

[0108] In actual implementation, the first ground image corresponding to all target geometric models in the first panoramic image coordinate system can be translated, and the height and width of the panoramic image to be filled are acquired based on the first pixel coordinate set of each first ground image.

[0109] In some embodiments, the target color value can be acquired based on the target geometric model corresponding to each first ground image and the coordinate of the first ground image in the first panoramic image coordinate system.

[0110] According to the ground panoramic image acquisition method, a plurality of ground images are acquired based on the target pan-tilt camera, a target geometric model of each ground image is acquired based on a geometric feature point pair between the ground image and an adjacent ground image, all target geometric models of the first panoramic image coordinate system are projected to the second panoramic image coordinate system, and a to-be-filled panoramic image is acquired; the to-be-filled image includes a first pixel coordinate set of each first ground image, the first pixel coordinate set includes a plurality of first pixel coordinate points, and each first pixel coordinate point is a coordinate of a pixel point of the first ground image in the second panoramic image coordinate system; for each first pixel coordinate set, a target color value of the first ground image is acquired based on the target geometric model; and the to-be-filled panoramic image is color-filled based on all target color values to acquire a ground panoramic image. By using the geometric feature point pair with high geometric splicing precision between each two adjacent ground images to generate the target geometric model of each ground image, the ground panoramic image is spliced and acquired, and compared with the panoramic image acquisition method based on texture features, the effect of acquiring the ground panoramic image for agricultural monitoring is significantly improved.

[0111] In some embodiments, all target geometric models of the first panoramic image coordinate system are projected to the second panoramic image coordinate system to acquire first pixel coordinate points of all pixel points of each first ground image in the second panoramic image coordinate system; a difference between a maximum value and a minimum value of the horizontal coordinates of the first pixel coordinate points in all first ground images is determined as a width of the to-be-filled panoramic image, and a difference between a maximum value and a minimum value of the vertical coordinates of the first pixel coordinate points in all first ground images is determined as a height of the to-be-filled panoramic image; and all first ground images in the second panoramic image coordinate system are framed based on the height and the width of the to-be-filled panoramic image to acquire the to-be-filled panoramic image.

[0112] In some embodiments, as shown in Figure 4 the difference between the maximum value and the minimum value of the horizontal coordinates of the first pixel coordinate points in all first ground images is determined as the width of the to-be-filled panoramic image, and the difference between the maximum value and the minimum value of the vertical coordinates of the first pixel coordinate points in all first ground images is determined as the height of the to-be-filled panoramic image, to acquire the to-be-filled panoramic image including all first ground images.

[0113] According to the ground panoramic image acquisition method, a plurality of ground images are acquired based on the target pan-tilt camera, a target geometric model of each ground image is acquired based on a geometric feature point pair between the ground image and an adjacent ground image of the ground image, panoramic image stitching is performed based on all target geometric models, and a ground panoramic image is acquired, so that the target geometric model of each ground image is generated based on the geometric feature point pair with high geometric stitching precision between each two adjacent ground images, and then the ground panoramic image is acquired by stitching, which significantly improves the effect of acquiring the ground panoramic image for agricultural monitoring compared with the way of acquiring the panoramic image based on texture features.

[0114] In some embodiments, a plurality of color values are acquired based on the target geometric model and the second pixel coordinate point; the second pixel coordinate point is a coordinate of a pixel corresponding to the first pixel coordinate point in the first panoramic image coordinate system; and the color values in the target coordinate range are determined as the target color values of the first ground image.

[0115] In some embodiments, the color values can be acquired based on the following formula:

[0116]

[0117] wherein Q i represents the color value of the i th first ground image, Pg represents a coordinate value of the second pixel coordinate point, represents an inverse matrix of the target geometric model corresponding to the i th first ground image.

[0118] In actual implementation, the target coordinate range can be a coordinate range of the first ground image in the second panoramic image coordinate system.

[0119] In some embodiments, after the plurality of color values of the i th first ground image are acquired, the color values that meet the target coordinate range of the i th first ground image are screened out as the target color values of the i th first ground image.

[0120] According to the ground panoramic image acquisition method, a plurality of ground images are acquired based on the target pan-tilt camera, a target geometric model of each ground image is acquired based on a geometric feature point pair between the ground image and an adjacent ground image of the ground image, panoramic image stitching is performed based on all target geometric models, and a ground panoramic image is acquired, so that the target geometric model of each ground image is generated based on the geometric feature point pair with high geometric stitching precision between each two adjacent ground images, and then the ground panoramic image is acquired by stitching, which significantly improves the effect of acquiring the ground panoramic image for agricultural monitoring compared with the way of acquiring the panoramic image based on texture features.

[0121] For better understanding of the ground panoramic image acquisition method provided by the embodiments of the present application, further explanation is made below, and it should be understood that the following discussion is only exemplary.

[0122] The present application provides a ground panoramic image acquisition method, and the specific steps can be as shown in Figure 5

[0123] Step 510: acquiring a plurality of ground images based on the target pan-tilt camera; the number of ground images is at least 2, and each ground image includes a plurality of geometric feature points.

[0124] In some embodiments, for the ground surface that needs to be monitored for agriculture, a data acquisition device can be installed on the top of a high tower on the ground surface, and the data acquisition device includes a high-definition camera combined with a pan-tilt (target pan-tilt camera), a GPS positioning and measuring device, etc. The high-definition camera can be controlled by the pan-tilt to rotate around the high tower to take pictures of the ground surface, and a plurality of ordered ground images are collected, each of which has a certain degree of overlap with the adjacent image.

[0125] In some embodiments, the target pan-tilt camera can be controlled to rotate horizontally by 60 degrees to obtain 6 ground images to be spliced.

[0126] In some embodiments, after acquiring a plurality of ground images, the plurality of ground images can be preprocessed.

[0127] In some embodiments, after acquiring a plurality of ground images, in the case that the ground resolutions of the plurality of ground images are different, a bilinear interpolation or a cubic convolution interpolation method can be used to resample the plurality of ground images to obtain a plurality of ground images with the same ground resolution.

[0128] In some embodiments, after acquiring a plurality of ground images, the geometric feature points of each ground image can be acquired.

[0129] In some embodiments, the geometric feature points of each ground image can be acquired based on a scale invariant feature transform (SIFT) algorithm, a Harris corner point algorithm, or any theoretically feasible algorithm.

[0130] Step 513: matching the geometric feature points in the first ground image and the geometric feature points in the second ground image to obtain a geometric feature point pair; the first ground image is any image in the plurality of ground images, and the second ground image is the adjacent image of the first ground image.

[0131] ​In some embodiments, after obtaining the geometric feature points of each surface image, the geometric feature points on two adjacent surface images can be matched for nearest neighbor points to obtain pairs of geometric feature points between the two adjacent surface images.

[0132] In practice, a geometric feature point pair can be a pair of matching geometric feature points in the overlapping area of ​​two adjacent surface images.

[0133] Step 516: If the number of geometric feature point pairs is greater than or equal to the feature point number threshold, obtain the first geometric model of the first surface image based on the homogeneous coordinates of the geometric feature point pairs.

[0134] In actual implementation, the feature point number threshold can be a set value. For example, the feature point number threshold can be set based on the number of geometric feature point pairs.

[0135] In some embodiments, when the number of geometric feature point pairs is greater than or equal to a feature point number threshold, a first geometric model of the first land surface image can be obtained based on the homogeneous coordinates of the geometric feature point pairs. For example, when the feature point number threshold is 10, the number of geometric feature point pairs is 12, and a first geometric model of the first land surface image can be obtained based on the homogeneous coordinates of the geometric feature point pairs.

[0136] In some embodiments, when the number of geometric feature point pairs is greater than or equal to a feature point number threshold, a first geometric model of the first surface image can be obtained based on the following formula:

[0137]

[0138] Among them, H ia Let x represent the first geometric model of the i-th first surface image, and let x be the geometric transformation between the geometric feature points of the first surface image and the geometric feature points of the second surface image. i1 =H ia x i2 x i1 Let x be the homogeneous coordinate of the geometric feature point of the i-th first surface image. i2 Let be the homogeneous coordinates of the geometric feature points of the second surface image corresponding to the i-th first surface image.

[0139] Step 519: When the number of geometric feature point pairs is less than the feature point number threshold, obtain the first geometric model of the first surface image based on the surface image resolution, the target coordinates of the first surface image, and the target coordinates of the second surface image.

[0140] In some embodiments, in a case where the number of geometric feature point pairs is less than the feature point number threshold, the first geometric model of the first ground surface image can be obtained based on the ground surface image resolution, the target coordinates of the first ground surface image, and the target coordinates of the second ground surface image. For example, in a case where the feature point number threshold is 10 and the number of geometric feature point pairs is 8, the first geometric model of the first ground surface image can be obtained based on the ground surface image resolution, the target coordinates of the first ground surface image, and the target coordinates of the second ground surface image.

[0141] In actual implementation, the target coordinates can be the starting ground coordinates of the first ground surface image or the second ground surface image, such as the coordinates of the center pixel point of the first ground surface image or the second ground surface image, or the coordinates of the top-left pixel point of the first ground surface image or the second ground surface image, or the coordinates of a pixel point at any position.

[0142] In some embodiments, in a case where the number of geometric feature point pairs is less than the feature point number threshold, the first geometric model of the first ground surface image can be obtained based on the following formula:

[0143]

[0144] t x =(X j -X i ) / GSD;

[0145] t y =(Y j -Y i ) / GSD;

[0146] wherein H ia represents the first geometric model of the i-th first ground surface image, (X i , Y i ) represents the target coordinates of the first ground surface image, (X j , Y j ) represents the target coordinates of the second ground surface image, and GSD is the ground resolution.

[0147] Step 522, projecting the first geometric model to the first panoramic image coordinate system to obtain a second geometric model of the first ground surface image; obtaining distance errors of all second geometric models; removing error feature points in the geometric feature points of each first ground surface image based on the distance errors by using the random sample consensus algorithm, and recalculating the second geometric model based on the geometric feature points after removing the error feature points; in a case where the distance error is less than or equal to an error threshold, determining the second geometric model corresponding to the distance error as a target geometric model.

[0148] In actual implementation, the first panoramic image coordinate system can be a coordinate system established with the target coordinate of any one first ground image as the coordinate system origin.

[0149] In some embodiments, the second geometric model of the first ground image can be obtained based on the following formula:

[0150] H1=I;

[0151] H i =H i-1 H ia ;

[0152] wherein I represents a unit matrix, H1 represents the second geometric model of the first first ground image, H i represents the second geometric model of the i-th first ground image, and H ia represents the first geometric model of the i-th first ground image.

[0153] In some embodiments, the distance error of all second geometric models can be obtained based on the following formula:

[0154]

[0155] wherein G represents the distance error of all second geometric models, H i represents the second geometric model of the i-th first ground image in the N first ground images, x i represents the horizontal coordinate of the target coordinate of the i-th first ground image, H i+1 represents the second geometric model of the second ground image corresponding to the i-th first ground image, and x i+1 represents the horizontal coordinate of the target coordinate of the second ground image corresponding to the i-th first ground image.

[0156] In some embodiments, the distance error can be used for nonlinear iterative optimization processing (Bundle Adjust), to remove error feature points in the geometric feature points of each first ground image, and the second geometric model can be recalculated based on the geometric feature points after removing the error feature points, and the distance error can be obtained based on the recalculated second geometric model. In the case where the distance error is greater than the error threshold, the above process is repeated. In the case where the distance error is less than or equal to the error threshold, the second geometric model corresponding to the distance error is determined as the target geometric model.

[0157] Step 525, project all target geometric models of the first panoramic image coordinate system to the second panoramic image coordinate system to obtain a first pixel coordinate point of all pixel points of each first ground image in the second panoramic image coordinate system; determine a difference between a maximum value and a minimum value of the horizontal coordinates of all first pixel coordinate points in all first ground images as a width of the to-be-filled panoramic image, and determine a difference between a maximum value and a minimum value of the vertical coordinates of all first pixel coordinate points in all first ground images as a height of the to-be-filled panoramic image; frame all first ground images in the second panoramic image coordinate system based on the height and the width of the to-be-filled panoramic image to obtain the to-be-filled panoramic image.

[0158] In actual implementation, the to-be-filled image includes a first pixel coordinate set of each first ground image, and the first pixel coordinate set includes a plurality of first pixel coordinate points, each first pixel coordinate point being a coordinate of a pixel point of the first ground image in the second panoramic image coordinate system.

[0159] In some embodiments, the difference between the maximum value and the minimum value of the horizontal coordinates of all first pixel coordinate points in all first ground images is determined as the width of the to-be-filled panoramic image, and the difference between the maximum value and the minimum value of the vertical coordinates of all first pixel coordinate points in all first ground images is determined as the height of the to-be-filled panoramic image, so as to obtain the to-be-filled panoramic image including all first ground images.

[0160] Step 528, based on the target geometric model and the second pixel coordinate point, obtain a plurality of color values; the second pixel coordinate point is a coordinate of a pixel point corresponding to the first pixel coordinate point in the first panoramic image coordinate system; and determine a color value in the target coordinate range as a target color value of the first ground image.

[0161] In some embodiments, the color value can be obtained based on the following formula:

[0162]

[0163] wherein Q i represents the color value for the i-th first ground image, Pg represents a coordinate value of the second pixel coordinate point, represents an inverse matrix of the target geometric model corresponding to the i-th first ground image.

[0164] In actual implementation, the target coordinate range can be a coordinate range of the first ground image in the second panoramic image coordinate system.

[0165] In some embodiments, after obtaining a plurality of color values for the i-th first ground image, the color value meeting the target coordinate range of the i-th first ground image is screened out as the target color value of the i-th first ground image.

[0166] In step 531, the color filling is performed on the to-be-filled panoramic image based on all target color values, and the ground panoramic image is obtained.

[0167] The embodiment of the present application further provides a ground panoramic image acquisition device.

[0168] In some embodiments, as shown in the figure, the ground panoramic image acquisition device 600 comprises a first acquisition module 610, a second acquisition module 620 and a third acquisition module 630. Figure 6

[0169] The first acquisition module 610 is configured to acquire a plurality of ground images based on a target pan-tilt camera; the number of ground images is at least 2, and each ground image comprises a plurality of geometric feature points.

[0170] The second acquisition module 620 is configured to acquire a target geometric model of a first ground image based on a geometric feature point pair between the first ground image and a second ground image; the first ground image is any image in the plurality of ground images, the second ground image is an adjacent image of the first ground image, and the geometric feature point pair is a geometric feature point pair matched in the first ground image and the second ground image.

[0171] The third acquisition module 630 is configured to perform panoramic image stitching based on all target geometric models to obtain a ground panoramic image.

[0172] According to the ground panoramic image acquisition device of the embodiment of the present application, a plurality of ground images are acquired based on a target pan-tilt camera, a target geometric model of each ground image is generated based on a geometric feature point pair between the ground image and its adjacent ground image, and a ground panoramic image is obtained by performing panoramic image stitching based on all target geometric models, so that the target geometric model of each ground image is generated based on the geometric feature point pair between each two adjacent ground images with high geometric stitching accuracy, and then the ground panoramic image is obtained by stitching, which significantly improves the effect of acquiring the ground panoramic image for agricultural monitoring compared with the method of acquiring the panoramic image based on texture features.

[0173] In some embodiments, the second acquisition module 620 comprises:

[0174] The first acquisition unit is configured to match the geometric feature points in the first ground image and the geometric feature points in the second ground image to obtain the geometric feature point pair.

[0175] The second acquisition unit is configured to acquire a first geometric model of the first ground image based on the number of geometric feature point pairs.

[0176] The third acquisition unit is configured to optimize the first geometric model by using a random sample consensus algorithm to obtain the target geometric model.​

[0177] In some embodiments, the second obtaining unit is configured to:

[0178] In a case where the number of the geometric feature point pairs is greater than or equal to the feature point number threshold, obtain the first geometric model of the first ground surface image based on the homogeneous coordinates of the geometric feature point pairs.

[0179] In some embodiments, the second obtaining unit is configured to:

[0180] In a case where the number of the geometric feature point pairs is less than the feature point number threshold, obtain the first geometric model of the first ground surface image based on the ground surface image resolution, the target coordinates of the first ground surface image, and the target coordinates of the second ground surface image.

[0181] In some embodiments, the third obtaining unit is configured to:

[0182] project the first geometric model to the first panoramic image coordinate system to obtain a second geometric model of the first ground surface image;

[0183] obtain distance errors of all the second geometric models;

[0184] remove error feature points in the geometric feature points of each first ground surface image based on the distance errors by using a random sample consensus algorithm, and recompute the second geometric model based on the geometric feature points after the error feature points are removed;

[0185] In a case where the distance error is less than or equal to an error threshold, determine the second geometric model corresponding to the distance error as a target geometric model.

[0186] In some embodiments, the third obtaining module 630 includes:

[0187] a fourth obtaining unit configured to project all the target geometric models of the first panoramic image coordinate system to a second panoramic image coordinate system to obtain a to-be-filled panoramic image; the to-be-filled image includes a first pixel coordinate set of each first ground surface image, and the first pixel coordinate set includes a plurality of first pixel coordinate points, each first pixel coordinate point being a coordinate of a pixel point of the first ground surface image in the second panoramic image coordinate system;

[0188] a fifth obtaining unit configured to, for each first pixel coordinate set, obtain a target color value of the first ground surface image based on the target geometric model;

[0189] a sixth obtaining unit configured to perform color filling on the to-be-filled panoramic image based on all the target color values to obtain a ground surface panoramic image.

[0190] In some embodiments, the fourth obtaining unit is configured to:

[0191] projecting all target geometric models of the first panoramic image coordinate system to the second panoramic image coordinate system to obtain a first pixel coordinate point of each pixel point of the first ground surface image in the second panoramic image coordinate system;

[0192] determining a difference between a maximum value and a minimum value of the horizontal coordinates of the first pixel coordinate points in all the first ground surface images as a width of the panoramic image to be filled, and determining a difference between a maximum value and a minimum value of the vertical coordinates of the first pixel coordinate points in all the first ground surface images as a height of the panoramic image to be filled;

[0193] based on the height and the width of the panoramic image to be filled, framing all the first ground surface images in the second panoramic image coordinate system to obtain the panoramic image to be filled.

[0194] In some embodiments, the fifth obtaining unit is configured to:

[0195] obtaining a plurality of color values based on the target geometric model and the second pixel coordinate point, the second pixel coordinate point being a coordinate of the pixel point corresponding to the first pixel coordinate point in the first panoramic image coordinate system;

[0196] determining the color values in the target coordinate range as the target color values of the first ground surface image.

[0197] The ground surface panoramic image obtaining apparatus in the embodiments of the present application can be an electronic device or a component in an electronic device, such as an integrated circuit or a chip. The electronic device can be a terminal or other devices other than a terminal. For example, the electronic device can be a mobile phone, a tablet computer, a notebook computer, a palm computer, a vehicle-mounted electronic device, a Mobile Internet Device (MID), an augmented reality (AR) / virtual reality (VR) device, a robot, a wearable device, an ultra-mobile personal computer (UMPC), a netbook, or a personal digital assistant (PDA), etc. The electronic device can also be a server, a Network Attached Storage (NAS), a personal computer (PC), a television (TV), a teller machine, or a self-service machine, etc. The embodiments of the present application are not limited in this regard.

[0198] The ground panoramic image acquisition device in the embodiments of the present application can be a device with an operating system. The operating system can be a Microsoft (Windows) operating system, an Android operating system, an IOS operating system, or other possible operating systems, and the embodiments of the present application are not limited in this regard.

[0199] The ground panoramic image acquisition device 600 provided in the embodiments of the present application can realize Figures 1 to 5 The method embodiments realize various processes, and to avoid repetition, the various processes are not described herein again.

[0200] In some embodiments, as shown in Figure 7 The embodiments of the present application also provide an electronic device 700, which includes a processor 701, a memory 702, and a computer program stored in the memory 702 and capable of running on the processor 701. When the program is executed by the processor 701, various processes of the above ground panoramic image acquisition method embodiments are realized, and the same technical effects are achieved. To avoid repetition, the various processes are not described herein again.

[0201] It should be noted that the computer device in the embodiments of the present application includes the mobile electronic device and the non-mobile electronic device described above.

[0202] The embodiments of the present application also provide a non-transitory computer readable storage medium, which stores a computer program. When the computer program is executed by a processor, various processes of the above ground panoramic image acquisition method embodiments are realized, and the same technical effects are achieved. To avoid repetition, the various processes are not described herein again.

[0203] The processor is the processor in the electronic device described in the above embodiments. The readable storage medium includes a computer readable storage medium, such as a computer readable memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0204] The embodiments of the present application also provide a computer program product, which includes a computer program. When the computer program is executed by a processor, the above ground panoramic image acquisition method is realized.

[0205] The processor is the processor in the electronic device described in the above embodiments. The readable storage medium includes a computer readable storage medium, such as a computer readable memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0206] The embodiment of the present application further provides a chip, which comprises a processor and a communication interface, the communication interface is coupled with the processor, the processor is used for running programs or instructions, realizes each process of the above-mentioned ground panoramic image acquisition method embodiment, and can achieve the same technical effects. To avoid repetition, it will not be repeated here.

[0207] It should be understood that the chip mentioned in the embodiment of the present application can also be referred to as a system chip, a system chip, a chip system or a system on chip, etc.

[0208] It should be noted that in this paper, the term "including", "containing" or any other variant thereof is intended to cover non-exclusive inclusion, so that the process, method, article or device including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such process, method, article or device. Without more limitations, the element defined by the statement "including a" does not exclude the presence of other identical elements in the process, method, article or device including the element. In addition, it should be pointed out that the scope of the method and device in the embodiment of the present application is not limited to the order of the functions shown or discussed, but also includes the functions performed in a substantially simultaneous manner or in the opposite order, for example, the described method can be performed in a different order from the described order, and various steps can also be added, omitted or combined. In addition, the features described with reference to some examples can be combined in other examples.

[0209] From the above description of the embodiments, those skilled in the art can clearly understand that the above-mentioned embodiment method can be realized by software and necessary general hardware platform, of course, it can also be realized by hardware, but in many cases, the former is a better embodiment. Based on such understanding, the technical solutions of the present application can be embodied in the form of a computer software product, which is stored in a storage medium (such as ROM / RAM, magnetic disc, optical disc), including a plurality of instructions for making a terminal (which can be a mobile phone, computer, server or network equipment, etc.) execute the method described in each embodiment of the present application.

[0210] The embodiments of the present application are described above in combination with the drawings, but the present application is not limited to the above-mentioned specific embodiments, the above-mentioned specific embodiments are only illustrative, not restrictive, and those skilled in the art can make many forms under the inspiration of the present application without departing from the scope of the present application and the protection scope of the claims.

[0211] In the description of the application, reference has been made to descriptive terms such as "one embodiment", "some embodiments", "an embodiment", "example", "specific example" or "some examples" etc. It is emphasized that each of these terms refers to a specific feature, structure, material or characteristic described in connection with a particular embodiment or example. The descriptive terms are not necessarily referring to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0212] While embodiments of the application have been shown and described, it is to be understood that the application is not limited to the details of the embodiments described, since the scope of the application is defined with respect to the appended claims.

Claims

1. A method for acquiring panoramic images of the earth's surface, characterized in that, include: Several surface images are acquired based on the target gimbal camera; the number of surface images is at least 2, and each surface image includes several geometric feature points; Based on the geometric feature point pairs between each first surface image and second surface image, the target geometric model of the first surface image is obtained; the first surface image is any image among the plurality of surface images, the second surface image is an adjacent image of the first surface image, and the geometric feature point pairs are the matching geometric feature points in the first surface image and the second surface image. A panoramic image of the land surface is obtained by performing panoramic image stitching based on all the target geometric models.

2. The method for acquiring panoramic surface images according to claim 1, characterized in that, The step of obtaining the target geometric model of the first surface image based on each pair of geometric feature points between the first and second surface images includes: The geometric feature points in the first surface image and the geometric feature points in the second surface image are matched to obtain geometric feature point pairs; Based on the number of geometric feature point pairs, a first geometric model of the first surface image is obtained; The first geometric model is optimized using a random sampling consensus algorithm to obtain the target geometric model.

3. The method for acquiring panoramic surface images according to claim 2, characterized in that, The step of obtaining the first geometric model of the first surface image based on the number of geometric feature point pairs includes: If the number of geometric feature point pairs is greater than or equal to the feature point number threshold, a first geometric model of the first surface image is obtained based on the homogeneous coordinates of the geometric feature point pairs.

4. The method for acquiring panoramic surface images according to claim 2, characterized in that, The step of obtaining the first geometric model of the first surface image based on the number of geometric feature point pairs includes: If the number of geometric feature point pairs is less than the feature point number threshold, a first geometric model of the first surface image is obtained based on the surface image resolution, the target coordinates of the first surface image, and the target coordinates of the second surface image.

5. The method for acquiring panoramic surface images according to claim 2, characterized in that, The step of optimizing the first geometric model using a random sampling consensus algorithm to obtain the target geometric model includes: The first geometric model is projected onto the first panoramic image coordinate system to obtain the second geometric model of the first surface image. Obtain the distance error for all second geometric models; Using a random sampling consensus algorithm, based on the distance error, error feature points are removed from the geometric feature points of each of the first surface images, and the second geometric model is recalculated based on the geometric feature points after removing the error feature points. If the distance error is less than or equal to the error threshold, the second geometric model corresponding to the distance error is determined as the target geometric model.

6. The method for acquiring panoramic surface images according to claim 1, characterized in that, The step of performing panoramic image stitching based on all the target geometric models to obtain a panoramic image of the land surface includes: Project all the target geometric models in the first panoramic image coordinate system onto the second panoramic image coordinate system to obtain a panoramic image to be filled; the panoramic image to be filled includes a set of first pixel coordinates for each of the first surface images, the set of first pixel coordinates includes a plurality of first pixel coordinate points, and each first pixel coordinate point is the coordinate of a pixel point of the first surface image in the second panoramic image coordinate system. For each set of first pixel coordinates, the target color value of the first surface image is obtained based on the target geometric model; Based on all the target color values, the panoramic image to be filled is color-filled to obtain a panoramic image of the land surface.

7. The method for acquiring panoramic surface images according to claim 6, characterized in that, The step of projecting all the target geometric models in the first panoramic image coordinate system to the second panoramic image coordinate system to obtain the panoramic image to be filled includes: Project all the target geometric models in the first panoramic image coordinate system to the second panoramic image coordinate system, and obtain the first pixel coordinates of all the pixels of each first surface image in the second panoramic image coordinate system. The difference between the maximum and minimum x-coordinates of the first pixel coordinates in all the first surface images is determined as the width of the panoramic image to be filled, and the difference between the maximum and minimum y-coordinates of the first pixel coordinates in all the first surface images is determined as the height of the panoramic image to be filled. Based on the height and width of the panoramic image to be filled, all the first surface images in the second panoramic image coordinate system are selected to obtain the panoramic image to be filled.

8. The method for acquiring panoramic surface images according to claim 6, characterized in that, The step of obtaining the target color value of the first surface image for each set of first pixel coordinates based on the target geometric model includes: Based on the target geometric model and the second pixel coordinate point, several color values ​​are obtained; the second pixel coordinate point is the coordinate of the pixel point corresponding to the first pixel coordinate point in the first panoramic image coordinate system. The color values ​​within the target coordinate range are determined as the target color values ​​of the first surface image.

9. A device for acquiring panoramic images of the earth's surface, characterized in that, include: The first acquisition module is used to acquire a number of surface images based on the target gimbal camera; the number of surface images is at least 2, and each surface image includes a number of geometric feature points; The second acquisition module is used to acquire the target geometric model of the first surface image based on the geometric feature point pairs between each first surface image and the second surface image; the first surface image is any one of the plurality of surface images, the second surface image is an adjacent image of the first surface image, and the geometric feature point pairs are the matching geometric feature points in the first surface image and the second surface image; The third acquisition module is used to perform panoramic image stitching operations based on all the target geometric models to acquire a panoramic image of the land surface.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory, characterized in that, The processor executes the computer program to implement the steps of the method for acquiring panoramic images of the land surface as described in any one of claims 1-8.

Images