A land surveying method based on aerial image analysis

Abstract

The present invention belongs to the technical field of land surveying, and specifically discloses a land surveying method based on aerial image analysis. The method performs a preliminary evaluation on the terrain features of a target area before aerial photography to determine the terrain type, and then flexibly adjusts the aerial photography position parameters according to the terrain type in actual aerial photography, so that the aerial image can better fit the terrain state of the target area, and can maximize its reliability and accuracy as a benchmark for subsequent land size measurement, thereby significantly reducing the problem of boundary extraction failure caused by terrain changes during boundary extraction, reducing measurement deviations, and improving the overall accuracy of land measurement. At the same time, after the land size is measured through aerial images, dynamic land size measurement based on the aerial photography cycle is added, and then the dynamic land size measurement results are used to evaluate the land size change, so that the change of land size can be discovered and quantified in time, providing a scientific basis for decision-making.

Description

Technical Field

[0001] The invention belongs to the technical field of land surveying, and specifically discloses a land surveying method based on aerial image analysis. Background Art

[0002] Modern urban planning requires high-precision land survey data to support refined design and management. Traditional surveying methods (such as total stations and GPS) are accurate but time-consuming and labor-intensive, especially when surveying large areas.

[0003] With the continuous advancement of UAV technology and remote sensing technology, UAVs have become an efficient surveying tool. With their high-efficiency surveying capabilities and applicability in complex or dangerous terrains, UAVs are now widely used in the field of land surveying and have become an efficient, safe and commonly used means in modern urban planning and land management.

[0004] In the prior art, there are already technical solutions for land surveying using drone aerial photography. For example, the Chinese invention patent with publication number CN116380023B proposes a land surveying and mapping system based on remote sensing technology. The system combines aerial photography technology and remote sensing mapping technology to compare remote sensing images according to their type characteristics, identify boundaries, and use these boundaries as the proposed flight route of the drone. Then, the image obtained by drone aerial photography is compared with the remote sensing mapping image to accurately determine the image position, and the aerial flight distance and image distance are measured based on the ratio of the overlapping parts of the two, thereby improving the accuracy of remote sensing mapping and solving the problem of unclear boundary recognition of remote sensing images.

[0005] Another example is that the Chinese invention patent with publication number CN115326023B proposes a land measurement preprocessing method based on drone images. The solution uploads land images taken by the drone's all-round high-definition camera to the drone image data processing platform for identification, and determines the corresponding area calculation of the land based on the identification results.

[0006] However, the above two technical solutions failed to fully consider the terrain type of the area when using drones to take aerial photos of the area to be measured. This deficiency leads to the inability to dynamically adjust the aerial photography parameters (such as flight altitude, angle, etc.) according to the terrain characteristics, and there is a hidden danger of parameter solidification. Specifically, due to the lack of adaptive optimization of terrain characteristics, aerial images may not be able to accurately reflect the actual terrain status of the area to be measured, thereby reducing its reliability and accuracy as a benchmark for subsequent land size measurements. In addition, when dealing with complex and undulating terrain, traditional boundary recognition algorithms may fail due to problems such as shadows, occlusions, and reflections caused by terrain changes. This not only affects the accurate extraction of boundaries, but also causes deviations in the overall measurement results, thereby weakening the overall accuracy of land measurement. Summary of the invention

[0007] To this end, one purpose of an embodiment of the present application is to provide a land surveying method based on aerial image analysis, which effectively solves the problems mentioned in the background technology by flexibly adjusting the aerial photography parameters according to the terrain type of the target area during the process of using drone aerial photography.

[0008] The purpose of the present invention can be achieved through the following technical solutions: A land surveying method based on aerial image analysis, comprising the following steps: Preliminary assessment of the terrain features of the target area to determine the terrain type, wherein the terrain type includes flat terrain and undulating terrain.

[0009] Several control points are deployed on the ground in the target area according to the terrain type.

[0010] The aerial photography position parameters are determined according to the terrain type of the target area, including the aerial photography position, aerial photography height and aerial photography angle.

[0011] The UAV is controlled to take aerial photos of the target area under the determined aerial photography position parameters, and images of the target area corresponding to different aerial photography positions are obtained.

[0012] The target area images corresponding to different aerial photography positions are pre-processed and registered using the arranged control points.

[0013] The target area images corresponding to different aerial photography positions after preprocessing and registration are spliced ​​and fused to form a topographic map of the target area.

[0014] Extract the features of objects from the topographic map of the target area, mark their locations, and draw land boundaries based on the features of objects.

[0015] The land size representation is calculated based on the drawn land boundary line, wherein the land size representation includes a geometric representation and a shape representation.

[0016] In combination with all the above technical solutions, the positive effects of the present invention are as follows: 1. The present invention conducts a preliminary assessment of the terrain characteristics of the target area before aerial photography to determine the terrain type, and then flexibly adjusts the aerial photography position parameters according to the terrain type in actual aerial photography, so that the aerial image can better fit the terrain state of the target area, and can maximize its reliability and accuracy as a benchmark for subsequent land size measurement, thereby significantly reducing the problem of boundary extraction failure caused by terrain changes during boundary extraction, reducing measurement deviations, and improving the overall accuracy of land measurement.

[0017] 2. After measuring the land size through aerial images, the present invention also adds dynamic land size measurement based on the aerial photography cycle, and then uses the dynamic land size measurement results to evaluate the land size change. On the one hand, regular aerial photography and measurement can provide the latest land size data to ensure the timeliness and accuracy of the information. On the other hand, by comparing data from different periods, changes in land size can be discovered and quantified in a timely manner, providing a scientific basis for decision-making. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The present invention is further described using the accompanying drawings, but the embodiments in the accompanying drawings do not constitute any limitation to the present invention. A person skilled in the art can obtain other drawings based on the following drawings without creative work.

[0019] Figure 1 The present invention is a flowchart of the steps for implementing the method.

[0020] Figure 2 It is a schematic diagram of the arrangement of control points under the five-point layout when the terrain type of the target area is flat terrain according to the present invention.

[0021] Figure 3 This is a schematic diagram of the aerial photography position when the terrain type of the target area is flat terrain in the present invention. DETAILED DESCRIPTION

[0022] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0023] See also Figure 1 As shown, the present invention proposes a land surveying method based on aerial image analysis, comprising the following steps: a preliminary assessment of the terrain features of the target area to determine the terrain type, wherein the terrain type includes flat terrain and undulating terrain.

[0024] Applied to the above scheme, the terrain type is determined by referring to the following process: a pre-flight of the target area is performed using a UAV to collect an overview image of the target area to generate a digital elevation model.

[0025] It should be added that a moderate flight altitude (such as 50-100 meters) should be selected during the pre-flight to ensure coverage of the entire target area, and the image resolution should be obtained when collecting the overview image, and compared with the set resolution threshold. If the resolution of the overview image at a certain flight altitude reaches the resolution threshold, the flight altitude is recorded. The digital elevation model generated based on the overview image can intuitively display the height changes of the terrain.

[0026] The slope of each pixel is calculated based on the generated digital elevation model, where the slope represents the inclination of the terrain surface at that point, usually in degrees or percentages, and the slopes of adjacent pixels are judged to be similar. If the slopes of adjacent pixels are similar, the adjacent pixels are classified as pixels with the same slope, thereby marking different slope areas in the generated digital elevation model.

[0027] Specifically, when calculating the slope of a pixel, for a given pixel , Represents the row index of the pixel. Represents the column index of the pixel whose slope It can be calculated by the following formula: ,in , , Indicates the elevation value, , Respectively represent the resolution of pixels in the x-axis and y-axis directions, , They represent the partial derivatives along the x-axis and y-axis respectively, reflecting the rate of change of the terrain in the east-west and north-south directions.

[0028] Applied to the above scheme, the slope similarity of adjacent pixel points can be judged by subtracting the slopes of the adjacent pixel points and comparing them with the similar slope difference. If the slope difference of a certain adjacent pixel point is less than or equal to the similar slope difference, the slope is judged to be similar, otherwise the slope is judged to be not similar.

[0029] The slopes of all pixels in different slope areas are averaged to obtain the representative slopes corresponding to the different slope areas, and the elevations of the different slope areas are obtained at the same time.

[0030] The representative slopes corresponding to different slope areas are compared, and the maximum representative slope, the minimum representative slope and the regional areas corresponding to the maximum representative slope and the minimum representative slope are extracted.

[0031] Using expressions Get the slope difference index of the target area , where , Respectively represent the maximum representative slope and the minimum representative slope. , They represent the area corresponding to the maximum representative slope and the minimum representative slope respectively.

[0032] It should be pointed out that the maximum representative slope mentioned above represents the steepest part of the area, and the minimum representative slope represents the flattest part of the area. The regional area corresponding to the maximum representative slope reflects the proportion of the steepest terrain, and the regional area corresponding to the minimum representative slope reflects the proportion of the flattest terrain. When conducting slope difference analysis, the maximum representative slope and the minimum representative slope and the regional areas corresponding to the maximum representative slope and the minimum representative slope are introduced, thereby avoiding analysis deviations caused by differences in terrain proportions.

[0033] The slope difference index of the target area is compared with the pre-configured critical slope difference index. For example, the critical slope difference index can be 0.3. If the slope difference index of the target area is less than the critical slope difference index, the terrain type of the target area is determined to be flat terrain. Otherwise, the terrain type of the target area is determined to be undulating terrain.

[0034] Several control points are deployed on the ground in the target area according to the terrain type.

[0035] It is important to know that the layout of control points provides precise geographic coordinates (such as longitude and latitude, elevation, etc.) for topographic surveying, ensuring that all measurement data are based on a unified benchmark. Control points can be used for data correction, which can eliminate instrument errors, human errors, and deviations caused by environmental factors, ensuring the reliability of measurement results. Generally speaking, the setting of control points should follow the following principles: First, uniform distribution: control points should be evenly distributed throughout the survey area, avoiding concentration in a local area, so as to ensure the global geo-registration accuracy. Second, sufficient number: it is usually recommended to set at least 3 control points. Third, easy to identify: control points should be selected at locations that are easily identifiable on the ground, such as road intersections, building corners, obvious landmarks, etc., so as to accurately find them in aerial images. Fourth, stability: control points should be set on stable objects to avoid position changes caused by wind, rain or other natural factors.

[0036] The specific layout process is as follows: when the terrain type of the target area is determined to be flat terrain, a regular grid layout is used to set the control points. For example, a control point is set at a certain distance (such as 100 meters or 200 meters) to form a uniform grid.

[0037] In the above layout example, assuming you are measuring a flat farmland, you can set a control point at each of the four corners and the center of the farmland to form a five-point layout, see Figure 2 As shown in the figure, if the farmland area is large, a few more control points can be added in the middle to form a denser grid layout.

[0038] When the terrain type of the target area is determined to be undulating terrain, the boundary lines of different slope areas are extracted from the slope areas divided by the digital elevation model of the target area, and then control points are arranged on the boundary lines of different slope areas. The principle followed in this arrangement is that in areas with large terrain undulations, control points should be set as much as possible in places with significant elevation changes, which can help better correct changes in terrain height.

[0039] Marking is implemented on the deployed control points, and positioning equipment is used to actually locate each control point and record its latitude and longitude coordinates.

[0040] It should be added that after the control points are set up, they need to be marked. Natural landmarks, such as large rocks, are usually selected. These landmarks are usually stable and not easy to move. If there are no suitable natural landmarks, some markers can be set up artificially as control points. The common practice is to spray a brightly colored cross mark on the ground, or place objects with obvious markings (such as reflectors, wooden stakes, etc.). Make sure these markers are clearly visible in the aerial images.

[0041] The invention reasonably sets control points according to the characteristics of different terrains, which is a key step to ensure accurate geo-registration of aerial images and improve measurement accuracy. In flat terrain, a regular grid layout can be used, while in areas with large terrain fluctuations, special attention should be paid to locations with significant elevation changes. By combining reasonable control point settings with high-precision GPS positioning and appropriate image processing technology, the accuracy and reliability of land measurement can be effectively improved.

[0042] The aerial photography position parameters are determined according to the terrain type of the target area, wherein the aerial photography position parameters specifically include the aerial photography position, aerial photography height and aerial photography angle.

[0043] It should be added that when taking aerial photos of a target area, it is usually necessary to set up multiple aerial photography positions. This is because the camera angle of a drone is limited. Even if it is equipped with a wide-angle or ultra-wide-angle lens, the effective area it can capture is limited. For larger target areas, a single aerial photography position cannot provide sufficient coverage, resulting in some areas not being captured.

[0044] In the manner in which the above scheme can be implemented, the aerial photography position is determined as follows: when the terrain type of the target area is determined to be flat terrain, the overview image of the target area is used to preliminarily extract the boundary of the target area using edge detection, and the boundary turning points are extracted using corner point detection.

[0045] The geometric center point of the target area is obtained by calculating the average coordinates of the boundary turning points.

[0046] Draw a minimum circumscribed circle that includes all boundary turning points, and draw a horizontal diameter line and a vertical diameter line inside the circumscribed circle.

[0047] Divide the horizontal diameter line and the vertical diameter line at equal distances to generate multiple division points.

[0048] The position of the segmentation point and the position of the boundary turning point are combined as the aerial photography position. Specifically, if the segmentation point and the boundary turning point overlap, the segmentation point and the boundary turning point are taken as one aerial photography position, and if the segmentation point and the boundary turning point do not overlap, the segmentation point and the boundary turning point are taken as two aerial photography positions.

[0049] It is important to understand that, see Figure 3 As shown in the figure, when the terrain type of the target area is flat terrain, by drawing horizontal and vertical diameter lines within the circumscribed circle and performing equidistant divisions, these division points can ensure that the entire target area is evenly covered to avoid shooting blind spots. The division points and the positions of the boundary turning points are combined as aerial photography positions, which further enhances the precise coverage of the boundary of the target area. Evenly distributed aerial photography points help ensure sufficient overlap between adjacent images, which is crucial for later image stitching, three-dimensional reconstruction and other processing, and can improve the accuracy and efficiency of data processing.

[0050] In a further implementation of the above scheme, the aerial photography altitude is determined as follows: when the terrain type of the target area is determined to be flat terrain, the flight altitude is determined based on the resolution of the overview image collected by the drone in pre-flight.

[0051] In a further implementation of the above scheme, the aerial photography angle is determined as follows: when the terrain type of the target area is determined to be flat terrain, the aerial photography angle adopts a vertical downward angle to ensure that all images are orthophotos.

[0052] When the terrain type of the target area is undulating terrain, the aerial photography position is determined as follows: when it is determined that the terrain type of the target area is undulating terrain, the center point position of each slope area divided by the digital elevation model of the target area is selected as the aerial photography position.

[0053] The above method can ensure that each slope area is fully covered, avoiding shooting blind spots caused by terrain changes. In addition, the optimal flight altitude and viewing angle of the drone on undulating terrain will change with the changes in the terrain. Selecting the center point of the slope area as the aerial photography position helps adjust the altitude and viewing angle of the drone to obtain the best image quality.

[0054] When the terrain type of the target area is undulating terrain, the aerial photography height is determined as follows: when it is determined that the terrain type of the target area is undulating terrain, the elevation of each slope area divided into each slope area according to the digital elevation model of the target area is combined with the set constant relative height to determine the aerial photography height of the aerial photography position belonging to each slope area.

[0055] When the terrain type is undulating, the drone's terrain-following flight mode can be used to keep the drone at a constant relative height to the ground, ensuring consistent relative height and image resolution throughout the entire survey area. This mode can automatically adjust the flight altitude according to changes in the terrain to ensure image consistency and quality.

[0056] When the terrain type of the target area is undulating terrain, the aerial photography angle is determined as follows: when it is determined that the terrain type of the target area is undulating terrain, the representative slope of each slope area in each slope area is divided according to the digital elevation model of the target area as the aerial photography angle of the aerial photography position to which each slope area belongs.

[0057] It should be understood that in the undulating terrain, the representative slope of each slope area is used as the aerial photography angle of the aerial photography position of each slope area. This is the use of oblique photography technology, which can enable the drone to better adapt to the changes in terrain and avoid the occlusion and shadow problems caused by the undulating terrain, thereby improving the quality and resolution of the image. In addition, oblique photography can capture more side information, especially in complex terrain such as steep slopes or valleys, which helps to generate more accurate orthophotos. For example, if the representative slope of a certain area is 30°, the tilt angle of the drone can be set to about 30° in this area to better capture the surface features.

[0058] The UAV is controlled to take aerial photos of the target area under the determined aerial photography position parameters, and images of the target area corresponding to different aerial photography positions are obtained.

[0059] The target area images corresponding to different aerial photography positions are preprocessed and registered using the arranged control points. The specific operations are as follows: the target area images corresponding to different aerial photography positions are preprocessed, wherein the preprocessing operations include radiation correction, geometric correction, etc., to eliminate the influence of factors such as atmospheric effects and lens distortion on the images.

[0060] It should be added that before registration, ensure that the metadata of each image (such as shooting time, camera parameters, etc.) is correctly saved for subsequent processing.

[0061] After the preprocessing operation, the previously deployed control points are found and marked in each target area image.

[0062] Associate the marked control point with the actual coordinates recorded previously for the control point.

[0063] Based on the actual coordinates of the control points and their coordinates in the target area image, a spatial transformation matrix (such as affine transformation, projection transformation, etc.) is calculated to correct the spatial position of the image, and the calculated spatial transformation matrix is ​​applied to each target area image to align the images in the geographic coordinate system.

[0064] The bundle adjustment technique is used to globally optimize the images. This technique further improves the registration accuracy by jointly optimizing the internal and external orientation elements (camera parameters, position, posture, etc.) of all images. Through multiple iterative adjustments, the errors between images are gradually reduced until the best registration effect is achieved.

[0065] The residual between the actual coordinates of each control point and the registered coordinates is calculated to evaluate the accuracy of the registration.

[0066] It is important to know that the residual refers to the difference between the actual coordinates and the registered coordinates.

[0067] Specifically, when evaluating the accuracy of the registration, the calculated residual is compared with the allowed registration residual. For example, the allowed registration residual is a pixel unit. Since there is more than one control point, the residuals of all control points can be statistically analyzed to obtain an overall registration accuracy assessment. Common statistical indicators include: average residual: the average value of the residuals of all control points; maximum residual: the largest residual among all control points; root mean square error: the square root of the mean of the sum of squares of the residuals of all control points. When the residual of the comprehensive evaluation is higher than the allowed registration residual, the registration accuracy is assessed to be unsatisfactory. At this time, the registration method needs to be adjusted, such as checking the currently used registration transformation model (such as affine transformation, similarity transformation, etc.), and trying to adjust the model parameters or select a more appropriate transformation model. For another example, a multi-resolution strategy (such as the pyramid method) is used to gradually perform registration from low resolution to high resolution to reduce the impact of local errors.

[0068] It is important to understand that after acquiring the target area image, it is very important to perform registration based on the control points. The main purpose of registration is to ensure that images from different sources or acquired at different times can be accurately aligned so that these images can be compared and analyzed in the same coordinate system.

[0069] The target area images corresponding to different aerial photography positions after preprocessing and registration are spliced ​​and fused to form a topographic map of the target area.

[0070] Applied to the above scheme, the target area images corresponding to different aerial photography positions after preprocessing and registration are stitched and fused as follows: the target area images corresponding to different aerial photography positions are stitched, and the overlapping areas are focused, and the quality indicators and control point distribution density of several overlapping images belonging to the overlapping areas are detected at the same time.

[0071] In the example of the above operation, it is assumed that there are 7 target area images obtained, among which target area image 1 and target area image 2 have overlapping area a, and target area image 3, target area image 5 and target area image 7 have overlapping area b, then the overlapping images belonging to the overlapping area a are target area image 1 and target area image 2, and the overlapping images belonging to the overlapping area b are target area image 3, target area image 5 and target area image 7.

[0072] The quality indicators mentioned above can be contrast, resolution, signal-to-noise ratio and color depth, where contrast refers to the difference in brightness between the brightest and darkest parts of an image. Images with high contrast usually have clearer details and better visual effects. Resolution refers to the number of pixels contained in an image. Higher resolution means more pixel information and can provide finer image details. Signal-to-noise ratio refers to the ratio of signal strength to noise strength. A high signal-to-noise ratio means that there is relatively more useful information (signal) in the image and less noise. Color depth refers to the number of colors that can be represented by each pixel in the image. Images with high color depth can display more color levels, making the image look more real and natural.

[0073] Furthermore, the control point distribution density can obtain the coverage area and the number of control points of the overlapping image to which the overlapped area belongs, and then divide the coverage area by the number of control points.

[0074] The quality index of several overlapping images in the overlapped area is used to perform quality scoring. The specific scoring formula is: , where , , , They represent the contrast, resolution, signal-to-noise ratio, and color depth of the overlapping images in the overlapping area. , , , They represent the sum of contrast, resolution, signal-to-noise ratio, and color depth of all overlapping images in the overlapping area. represents a natural constant, Indicates the rating system. The value can be 1, 10, 100. When the value is 1, it means the score is one point. When the value is 10, it means the score is 10 points. A value of 100 indicates a 100-point rating.

[0075] It should be understood that the weights of contrast, resolution, signal-to-noise ratio and color depth in the above scoring formula are 0.2, 0.3, 0.3 and 0.2 respectively. The purpose of this is that although contrast is important for visual effects and certain specific applications, it has less impact on the overall image quality than other indicators. Resolution directly affects the detail performance and data integrity of the image, and is one of the key factors determining image quality. Signal-to-noise ratio directly affects the data reliability and clarity of the image, and is one of the key factors determining image quality. Color depth is very important for some applications, but it is not a decisive factor in many other applications.

[0076] The quality scores of each overlapping image in the overlapped area are combined with the control point distribution density to evaluate the selection value of each overlapping image in the overlapped area. The specific evaluation formula is: , where Indicates the selection value. Indicates quality rating. Indicates a full score. represents the distribution density of control points, Represents the sum of the distribution densities of control points in several overlapping images belonging to the overlapping area.

[0077] The above The value is 1, the full score is 1, when When the value is 10, the full score is 10. When the value is 100, the full score is 100.

[0078] From the selection values ​​of the overlapping images belonging to the overlapping area, the overlapping image corresponding to the maximum selection value is selected as the valid overlapping image, thereby retaining the overlapping area corresponding to the valid overlapping image and using it for the final topographic map generation.

[0079] The present invention performs comprehensive screening by combining image quality and control point distribution density when screening effective overlapping images from overlapping images belonging to the overlapping area, wherein a high-density control point distribution usually means a higher registration accuracy. This method not only improves the registration accuracy, but also enhances data consistency and reliability.

[0080] Extract geographical features from the topographic map of the target area. Geographic features such as buildings, roads, rivers, etc. are usually the most prominent and stable elements on the topographic map. They provide reliable geographic reference points for drawing land boundaries. These features can be used to more accurately mark the boundaries of different land use types, mark the positions of the extracted geographical features, and draw land boundary lines based on the geographical features. The specific process is: mark the extracted geographical features and record their geographic coordinates.

[0081] Use the line segment connection tool to draw land boundary lines based on the marked land features.

[0082] It should be pointed out that the land boundary lines drawn based on land features follow certain rules and standards, such as the corners of buildings, the center lines of roads, the shorelines of rivers, etc., so as to ensure the consistency and comparability of the data.

[0083] The land dimension representation is calculated based on the drawn land boundary line, wherein the land dimension representation includes geometric representation and shape representation, wherein the geometric representation may be perimeter and area, and the shape representation may be circularity, compactness and elongation.

[0084] The circularity mentioned above is an indicator to measure the degree to which a polygon is close to a circle. The higher the circularity, the closer the polygon is to a circle. The calculation formula of circularity is: , where Indicates area, Represents the circumference. For a perfect circle, the circularity is equal to 1, and for other shapes the circularity is less than 1.

[0085] Compactness is another indicator to measure the compactness of polygon shapes. The higher the compactness, the closer the polygon is to a circle. The calculation formula for compactness is: .

[0086] Elongation is an indicator that measures the degree of stretching of a polygon along its main axis. The higher the elongation, the more slender the polygon is. The elongation can be obtained by calculating the aspect ratio of the minimum circumscribed rectangle. An elongation greater than 1 indicates that the polygon is relatively slender, and an elongation equal to 1 indicates that the polygon is close to a square.

[0087] Furthermore, after calculating the land size representation based on the drawn land boundary line, it also includes setting an aerial photography cycle, and then continuing to perform aerial photography calculation of the land size representation of the target area at the same time in the subsequent aerial photography cycle, and comparing it with the land size representation calculated by the current aerial photography to evaluate whether the land size has changed, and predicting the direction of the change when it is evaluated that the land size has changed. The specific implementation is as follows: subtract the land size representations obtained from adjacent aerial photography cycles item by item, take the absolute value, and divide it by the interval length of the aerial photography cycle to obtain the change rate of each land size representation.

[0088] The change rate of each land size representation is compared with a preset critical change rate. For example, the critical change rate is 0.1. If the change rate of a land size representation reaches the critical change rate, it is assessed that the land size has changed.

[0089] The change rates of the land dimensions that have changed are compared, and the representation item with the maximum change rate is taken as the change direction.

[0090] After measuring the land size through aerial images, the present invention also adds dynamic land size measurement based on the aerial photography cycle, and then uses the dynamic land size measurement results to evaluate the land size change. On the one hand, regular aerial photography and measurement can provide the latest land size data to ensure the timeliness and accuracy of the information. On the other hand, by comparing data from different periods, changes in land size can be discovered and quantified in a timely manner, providing a scientific basis for decision-making.

[0091] The above contents are merely examples and explanations of the structure of the present invention. The technicians in this technical field may make various modifications or additions to the specific embodiments described or replace them in a similar manner. As long as they do not deviate from the structure of the invention or exceed the scope defined by the present invention, they should all fall within the protection scope of the present invention.

Claims

1. A land surveying method based on aerial image analysis, characterized in that: The following steps are involved: Conduct a preliminary assessment of the terrain characteristics of the target area to determine the terrain type, which includes flat terrain and undulating terrain; Deploy several control points on the ground in the target area according to the terrain type; Determine the aerial photography location parameters according to the terrain type of the target area, including the aerial photography location, aerial photography height and aerial photography angle; For flat terrain, the overview image of the target area is used to preliminarily extract the boundary of the target area using edge detection, and the boundary turning points are extracted using corner detection; The geometric center point of the target area is obtained by calculating the average value of the coordinates of the boundary turning points; Draw the minimum circumscribed circle containing all boundary turning points, and draw a horizontal diameter line and a vertical diameter line inside the circumscribed circle; Divide the horizontal diameter line and the vertical diameter line at equal distances to generate multiple segmentation points; The position of the segmentation point and the position of the boundary turning point are combined as the aerial photography position; For flat terrain, the flight altitude is determined based on the resolution of the overview image collected by the drone during pre-flight, and the aerial photography angle is vertically downward; For undulating terrain, the center point of each slope area divided by the digital elevation model of the target area is selected as the aerial photography position; Divide the elevation of each slope area in each slope area according to the digital elevation model of the target area, and determine the aerial photography height of the aerial photography position to which each slope area belongs in combination with the set constant relative height; The representative slope of each slope area is divided into each slope area according to the digital elevation model of the target area as the aerial photography angle of the aerial photography position to which each slope area belongs; Control the UAV to take aerial photos of the target area under the determined aerial photography position parameters, and obtain images of the target area corresponding to different aerial photography positions; The target area images corresponding to different aerial photography positions are pre-processed and registered using the arranged control points; The target area images corresponding to different aerial photography positions after preprocessing and registration are stitched and fused to form a topographic map of the target area; Extract the features of the objects from the topographic map of the target area, mark their locations, and draw the land boundary lines based on the features of the objects; The land size representation is calculated based on the drawn land boundary line, wherein the land size representation includes a geometric representation and a shape representation.

2. A land surveying method based on aerial image analysis as claimed in claim 1, characterized in that: The process of determining the terrain type is as follows: Use UAV to pre-fly the target area to collect overview images of the target area and generate a digital elevation model; The slope of each pixel is calculated based on the generated digital elevation model, and the slopes of adjacent pixels are judged to be similar. If the slopes of adjacent pixels are similar, the adjacent pixels are classified as pixels with the same slope, thereby marking different slope areas in the generated digital elevation model; The slopes of all pixels in different slope areas are averaged to obtain the representative slopes corresponding to the different slope areas, and the elevations of the different slope areas are obtained at the same time; Compare the representative slopes corresponding to different slope areas, and extract the maximum representative slope, the minimum representative slope, and the area corresponding to the maximum representative slope and the minimum representative slope; Using expressions Get the slope difference index of the target area , where , Respectively represent the maximum slope and the minimum slope. , They represent the area corresponding to the maximum representative slope and the minimum representative slope respectively; The slope difference index of the target area is compared with the pre-set critical slope difference index. If the slope difference index of the target area is less than the critical slope difference index, the terrain type of the target area is determined to be flat terrain, otherwise the terrain type of the target area is determined to be undulating terrain.

3. A land surveying method based on aerial image analysis as claimed in claim 2, characterized in that: The process of laying out several control points on the ground of the target area according to the terrain type is as follows: When the terrain type of the target area is determined to be flat terrain, a regular grid layout is used to set control points; When the terrain type of the target area is determined to be undulating terrain, the boundary lines of different slope areas are extracted from the slope areas divided by the digital elevation model of the target area, and then control points are arranged on the boundary lines of different slope areas; Marking is implemented on the deployed control points, and positioning equipment is used to actually locate each control point and record its latitude and longitude coordinates.

4. The land surveying method based on aerial image analysis according to claim 1, characterized in that: If the segmentation point and the boundary turning point overlap, the segmentation point and the boundary turning point are regarded as one aerial photography position. If the segmentation point and the boundary turning point do not overlap, the segmentation point and the boundary turning point are regarded as two aerial photography positions.

5. The land surveying method based on aerial image analysis according to claim 3, characterized in that: The target area images corresponding to different aerial photography positions are pre-processed and registered using the arranged control points, see the following process: Perform preprocessing operations on target area images corresponding to different aerial photography positions; Find and mark previously placed control points in each target area image after preprocessing operations; Associating the marked control points with the actual coordinates of the control points previously recorded; Calculate a spatial transformation matrix based on the actual coordinates of the control points and their coordinates in the target area image, and apply the calculated spatial transformation matrix to each target area image to align the images in the geographic coordinate system; Global optimization of the target area image is performed through bundle adjustment technology; The residual between the actual coordinates of each control point and the registered coordinates is calculated to evaluate the accuracy of the registration.

6. The land surveying method based on aerial image analysis according to claim 1, characterized in that: The process of splicing and fusing the target area images corresponding to different aerial photography positions after preprocessing and registration to form a topographic map of the target area is as follows: The target area images corresponding to different aerial photography positions are stitched together, and the overlapping areas are focused on. At the same time, the quality indicators and control point distribution density of several overlapping images belonging to the overlapping areas are detected; Quality scoring is performed using quality indicators of several overlapping images belonging to the overlapping area; The specific scoring formula is , where , , , They represent the contrast, resolution, signal-to-noise ratio, and color depth of the overlapping images in the overlapping area. , , , They represent the sum of contrast, resolution, signal-to-noise ratio, and color depth of all overlapping images in the overlapping area. represents a natural constant, Indicates the rating system; The quality scores of each overlapping image in the overlapped area are combined with the control point distribution density to evaluate the selection value of each overlapping image in the overlapped area. The specific evaluation formula is: , where Indicates the selection value. Indicates quality rating. Indicates a full score. represents the distribution density of control points, Represents the sum of the distribution densities of control points in several overlapping images belonging to the overlapping area; From the selection values ​​of the overlapping images belonging to the overlapping area, the overlapping image corresponding to the maximum selection value is selected as the valid overlapping image, thereby retaining the overlapping area corresponding to the valid overlapping image and using it for the final topographic map generation.

7. The land surveying method based on aerial image analysis according to claim 1, characterized in that: The process of drawing land boundary lines based on land features is as follows: Mark the extracted features and record their geographic coordinates; Use the line segment connection tool to draw land boundary lines based on the marked land features.

8. The land surveying method based on aerial image analysis according to claim 1, characterized in that: After calculating the land size representation based on the drawn land boundary line, the method further includes setting an aerial photography cycle, and then continuing to perform aerial photography calculation of the land size representation of the target area at the same time of the subsequent aerial photography cycle, and comparing with the land size representation calculated by the current aerial photography to evaluate whether the land size has changed, and predicting the direction of the change when it is evaluated that the land size has changed. The specific implementation is as follows: The land size representations obtained from adjacent aerial photography cycles are subtracted one by one, and the absolute value is divided by the interval between the aerial photography cycles to obtain the change rate of each land size representation; Compare the change rate of each land dimension representation with the preset critical change rate. If the change rate of a land dimension representation reaches the preset critical change rate, it is assessed that there is a change in the land dimension; The change rates of the land dimensions that have changed are compared, and the representation item with the maximum change rate is taken as the change direction.

Images