TW201938991A - Route planning method for aerial photography utilizing multi-axial unmanned aerial vehicle regulating the aerial photography height according to topography - Google Patents

Abstract

This invention relates to a route planning method for aerial photography utilizing a multi-axial unmanned aerial vehicle. The route planning method includes the steps of judging a topography of an aerial photography region to be a flat land or hilly land. If the aerial photography region is the hilly land, execute a hilly land airline planning mode. The hilly land airline planning mode includes the steps of judging a first hilly airline to be a straight line or an isohypse; calculating the number of hilly airlines; calculating an initial elevation value; and judging whether every two adjacent hilly airlines have a baseline lateral overlap rate conforming to a preset value or not. If the aerial photography region is the flat land, execute a flat land airline planning mode to generate a flat land airline route. The aerial photography height can also be regulated according to topographic relief on the premise that the standard lateral overlap rate and front-back overlap rate are met. The route planning method is suitable for the hilly land or a flat land with a complex height, and meanwhile, the resolution ratio of a picture is increased.

Description

Route planning method for aerial photography using multi-axis unmanned aerial vehicle

A route planning method for aerial photography, especially a route planning method for aerial photography using a multi-axis unmanned aerial vehicle.

In recent years, with the convenience and popularity of aerial cameras, aerial photogrammetry has become a hot industry. Aerial photogrammetry has a wide range of applications, from environmental disaster monitoring such as earthquakes, volcanic eruptions, and floods, to environmental monitoring such as the cover of planted trees, changes in river flows, and the production and planning of transportation routes in cities, and road detection. Both can use aerial photography technology of aerial cameras to obtain a large range of top views, it is easier to see the trend change in this range, and further develop appropriate policies.

Generally speaking, when taking aerial photogrammetry, no matter how the terrain of the aerial area to be photographed changes, the aerial camera will maintain a fixed aerial height, but aerial photogrammetry will face the following problems:

I. Inconsistent front-to-back overlap rates: See Figure 7. Assume that aerial camera 71 flies in the direction of rising terrain and takes a photo of the ground at a fixed interval. Numbers 1 to 8 represent the front and back of each photo The range, the numbers a, b, and c indicate the front-to-back overlapping area of two adjacent ground-to-ground photos. After superimposing and stitching each ground-to-ground photo, a ground-to-ground photo of one of the routes in the aerial area can be obtained. However, during the elevation of the terrain in the aerial photography area, the ground-level height of the aerial camera 71 remains unchanged, so that the overlapping area of two adjacent ground-level photos decreases as the terrain rises, as shown in the photo. The area b overlapping between numbers 4 and 5 is smaller than the area a overlapping between photo numbers 1 and 2. The area c overlapping between photo numbers 7 and 8 is smaller than the overlapping area b. As a whole, c <b <a, As a result, when the photos are stitched together, there will be an inconsistency in the overlap rate.

Second, the lateral overlap rate is inconsistent: see FIG. 8, the aerial camera 71 flies back and forth along the vertical direction of rising terrain. For example, photo number 1 is taken when the aerial camera 71 is flying in the direction above the picture. Photo number 2 It is taken when the aerial camera 71 is flying in the direction below the picture. Numbers 1 to 8 represent the range covered by each photo on the ground, path1 is the first path, and it represents the path when the aerial camera 71 took the photo number 1, path2. Is the second path, which represents the path when the aerial camera 71 takes the photo number 2, and so on. The drawing only uses the first path path1 to the fifth path path5 as examples, and the labels d, e, and f indicate two adjacent photos. The lateral overlap area of the photo. Similarly, because the terrain is getting higher and higher, but the flying height of the aerial camera 71 remains the same, so that during the shooting of the flight, the higher the altitude of the terrain, the range of the ground-based photos taken will be reduced, such as As shown in Figure 8, the range of photo number 1 is greater than photo number 2, and the range of photo number 2 is greater than photo number 3, which decreases in order. As the terrain rises, the range of photos acquired on the ground becomes smaller and smaller. , Causing the overlapping range between photos to decrease. As shown in the figure, as the terrain rises, the lateral overlap area e between photo numbers 4 and 5 is smaller than the lateral overlap area d of photo numbers 1 and 2, and photo number 7 The lateral overlap area f between 8 and 8 is smaller than the lateral overlap area e. On the whole, f <e <d, which causes the problem of inconsistent lateral overlap rates when the photos are stitched together.

Third, the resolutions are inconsistent: see FIG. 9, the reference g represents the difference in height of one location photographed by the aerial camera 71, and the reference i represents the difference in height of another location photographed by the aerial camera 71. As shown in the figure, the height difference g is greater than the height difference i, and the aerial height of the aerial camera 71 is maintained at a fixed horizontal level. However, the focal length of the camera used by the aerial camera 71 is a fixed value. The resolution is not the same as the resolution taken at the label i. As a result, in the same ground photograph, some areas can be clearly imaged, and some areas may be blurred, and the terrain situation of the area cannot be even judged.

In the original aerial route planning, because the terrain changed too much, but under the condition that the aerial height remains unchanged, the front-to-back overlap ratio and lateral overlap ratio between photos may not meet the standards, or the resolution of the photos taken It is not high enough to clearly show the undulations of the landform, resulting in the problem of uneven proportions and distortion of the photos.

In view of the previous aerial photography technology, the front-to-back overlap ratio and lateral overlap ratio of each aerial photo are inconsistent, and the resolution of each aerial photo is different, resulting in distortion, broken images and different scales of the map model produced. , This creative department proposes a route planning method for aerial photography using a multi-axis unmanned aerial vehicle. First, set the values of the front and back overlap ratio and lateral overlap ratio to plan the appropriate aerial photography path, and it can be adjusted according to the terrain. Adjusting the aerial height of the aerial camera can effectively solve the problem of inconsistent front-back overlap ratio, lateral overlap ratio and resolution, and provide more accurate and detailed aerial photos.

In order to achieve the above purpose, this creative route planning method using multi-axis unmanned aerial vehicles for aerial photography includes: (a) determining the terrain of an aerial photography area as flat or hillside, according to a maximum altitude value within the aerial photography area and A first height difference between a minimum altitude value determines whether the terrain of the aerial photography area is flat or hillside; when the aerial photography area is judged to be hillside, a hillside route planning mode is executed to generate one of the aerial photography areas Hillside route path. The hillside route path includes a plurality of hillside routes. Among them, the hillside route planning model involves the following steps: (b1) determining a first hillside route as a straight line or a contour line; (b2) calculating such hillside routes. Number, where the distance between each route is the distance between two adjacent hillside routes; when the aerial photography area is judged to be flat ground, a flatland route planning mode is executed to generate a flatland route path in the aerial photography area, and the flatland route path contains a plurality of Flat ground routes.

By using the pre-set front-back overlap ratio and lateral overlap ratio, combined with a numerical elevation model, no matter whether the terrain of the aerial photography area is flat or hillside, you can judge and plan the most suitable aerial photography path based on the relief of the terrain. Change the height of the camera, adjust the aerial height, keep the resolution of each photo within a certain range, and provide more complete and higher-resolution aerial photos.

Please refer to FIG. 1, which is a flowchart of the steps of this creation. This creation proposes a route planning method for aerial photography using a multi-axis unmanned aerial vehicle, in which a rotor-type unmanned aerial vehicle can achieve better aerial photography effects. This creation can be pre-set by the user on the aerial camera, including freely setting camera parameters, drawing parameters, and range coordinates of an aerial area, where the camera parameters include the focal length, the size of the photosensitive element, and the image resolution. The drawing parameters Includes ground resolution (GSD), front-to-back overlap ratio A%, lateral overlap ratio B%, and image measurement accuracy. The range coordinates of the aerial photography area are provided by a numerical elevation model. This embodiment uses ASTER 30m numerical elevation. mode. From the above data, values such as aerial height, baseline length b, and route distance Di can be obtained through calculation.

For example, first set the camera focal length 22mm, the sensor size 22.3 * 14.9mm (w * h), the image resolution 5184 * 3456 (w * h), and the ground resolution 1cm / Pixel in the mapping parameters. The front and back overlap ratio is 70%, the lateral overlap ratio is 50%, the image measurement accuracy is 10mm, and the aerial photography area range is 400 * 300m ² . The aerial photography height is calculated as: From the above data, the aerial height is 60m. The aerial photography height refers to the altitude of the aerial camera relative to the corresponding ground surface during the shooting operation. The calculation of the baseline length b is: From the above data, it can be obtained that the baseline length b is 12 m. Wherein, the baseline length b is the distance between the image centers of two adjacent photos on the ground. The route distance Di is calculated as:

Please refer to FIG. 2 at the same time. From the above data, the route distance Di is 30m, and the route distance Di is the distance between two adjacent hillside routes 111. In step S101, first determine whether the terrain of an aerial photography area 10 is a hillside or flat land. Among them, the terrain covered by the aerial photography area 10 is known to have a highest altitude value and a minimum altitude value. When the terrain of 10 is hillside or flat, the difference between the highest altitude value and the lowest altitude value is first calculated as a first height difference value Δh1, and the first height difference value Δh1 is divided by a set value A calculated value is obtained. In a preferred embodiment, the set value is 640. The calculated value is compared with a preset ground resolution value (GSD). If the calculated value is greater than the ground resolution Value, the aerial photography area 10 is judged to be a hillside. If the calculated value is smaller than the ground resolution, the aerial photography area 10 is judged to be a flat ground; the highest altitude value and the lowest altitude value can be obtained by the numerical elevation model. Following the above example, in the aerial photography area 10, the highest altitude value is 490m and the lowest altitude value is 330m. The first height difference value Δh1 is 160m. Dividing 160 by 640 results in a calculated value equal to 25cm. Since the calculated value (25 cm) is larger than the ground resolution value (1 cm), it can be determined that the aerial photography area 10 is a hillside land.

When the terrain of the aerial photography area 10 is judged as a hillside, a hillside route planning mode is executed, and the hillside planning mode is used to generate a hillside route route 11, wherein the hillside route route 11 includes a plurality of hillside routes 111 ; The hillside route route 11 is shown in FIG. 2, which represents an aerial camera 12 flying in the direction of a route (as indicated by the Y direction shown in FIG. 2) and following a shooting direction (as shown in FIG. 2). X direction) move back and forth, and take a ground-level photo every flight of a baseline length b, where the baseline length b is the distance between the image centers of two adjacent ground-level photos; the shooting direction is perpendicular to the terrain direction, and the route direction Run parallel to the terrain and fly from a lower terrain to a higher terrain; this hillside route planning model performs the following steps:

S102: Determine whether a first hillside route 21 is a straight line or a contour line. Please refer to FIG. 3. Firstly, the terrain of one of the boundaries of an aerial photography area 20 is selected as a surface profile, and the route corresponding to the surface profile is defined as the first hillside route 21, and the boundary is within the aerial photography area 20. The boundary near the lowest elevation. Then assume that the first hillside course 21 is a straight line, and calculate one of the second height difference values Δh2 of the first hillside route 21, where the second height difference value Δh2 is the highest point altitude value and The difference in altitude at the lowest point, the second height difference value Δh2 can be obtained by a high-level surface numerical model; then the second height difference value Δh2 is compared with a route threshold D, which is determined by a depth axis The accuracy formula is calculated, and the depth axis accuracy formula is: σ _z is the ground resolution value, D is the threshold value of the route, b is the baseline length, d is the camera focal length, and σ _i is the image measurement accuracy, where the ground resolution value σ _z , the baseline length b, and the camera focal length d And the image measurement accuracy σ _i is a preset known parameter. When the second height difference value Δh2 is less than the route threshold value D, the first hillside route 21 is determined to be a straight line; when the second height difference value △ h2 is greater than the route threshold value D, the first route is determined The slope route 21 is a contour line.

Following the above example, the ground resolution value is 1cm / Pixel, the preset baseline length b is 12m, the camera focal length is 22mm, and the image measurement accuracy is 10mm. According to the depth axis formula, the threshold value of the route D = 0.51m can be obtained. According to the numerical elevation model, it can be known that the highest elevation value of the surface profile is 342m and the lowest elevation value is 330m, so that the second height difference value Δh2 is equal to 12m. Because the second height difference value Δh2 (12m) is greater than the route threshold value D (0.51m), it is determined that the first hillside route 21 is a contour line.

S103: Calculate the number of the hillside routes 111 by using the calculated route distance Di. Use the following steps to get the number of these hillside routes 111: Project the aerial photography area 10 into a plane, and divide the distance of the route direction of the aerial photography area by the distance between the routes and add 1 to calculate the number of these hillside routes 111 .

Following the above example, since the direction of the route is 400m and the distance between the routes is 30m, 400m / 30m + 1 = 14, the number of such hillside routes 111 is 14.

Next, the route distance Di is corrected, and the correction steps are as follows: S104: Divide the first height difference value Δh1 by the number of the hillside routes 111 to obtain the initial height values of the corresponding contour lines between two adjacent route paths. The elevation value is the height difference between two adjacent contour lines; the height of the first contour line is sequentially added to the initial elevation value, and the heights of the second and third contour lines can be obtained. At the same time, since the aerial camera 12 flies The altitude of the ground route is the same as that of the corresponding ground, so that the first hillside course 21 is sequentially added with the initial elevation value, and the aerial height of each hillside course 111 can be obtained. For example, the first height difference value Δh1 is 160m, the total number of hillside routes 111 in aerial area 10 is 14, the initial elevation value is 160/14 = 11.4m, and the aerial height of the first hillside route 111 is 330m, the aerial height of the second hillside route 21 is 330 + 11.4 = 341.4m, the aerial height of the third hillside route is 352.8m, and so on, the aerial height of 111 for each hillside route can be obtained by analogy;

S105: Determine whether the lateral overlap rate of two adjacent hillside routes 111 meets the preset reference lateral overlap rate B%; as shown in FIG. 4, the distance between two adjacent hillside routes 111 may be different. The maximum distance between the 111 segments of two adjacent hillside routes may be smaller than the route distance Di, or it may be greater than the route distance Di. Using the nearest neighbor distance detection method, if the maximum distance between partial segments of two adjacent hillside routes 111 is greater than the calculated route distance Di, step S106 is performed to reduce the aerial height of the next hillside route 111 by an altitude calculation value. , So that the actual distance between two adjacent hillside routes 111 is not greater than the distance between the routes Di; in a preferred embodiment, the calculated value of the elevation can be obtained by iteration, and by repeating iteration, it is reduced by 1 meter each time. , In order to reduce the elevation calculation value in order, so that the actual distance between all the line segments in two adjacent hillside routes 111 is not greater than the route distance Di.

Following the above example, as shown in FIG. 5, the route interval Di is 30m, and the height of the first hillside route 21 is 330m. Assuming the original elevation calculation value is 50m, and adding 330m to the original elevation calculation value of 50m, a The height of the second hillside route 22 is 380m, but the original elevation calculation value is 50m, which is greater than the route distance Di, then the original altitude calculation value is reduced to 49m, and then compared with the route distance Di, if it is still greater than the route distance Di, then reduce the calculated height to 48m, then reduce it by 1m and compare it with the distance between the routes Di; when the calculated value is less than or equal to 30m, add the corrected altitude calculation value to the first slope route 21 30m, a second hillside correction route 23 is obtained, and all distances between the second hillside correction route 23 and the first hillside route 21 are not greater than the route distance Di, and the number of these hillside routes 111 is recalculated, and Repeatedly check whether the maximum distance between two adjacent hillside routes exceeds the distance between the routes Di, so as to ensure that the superimposed photos can meet the reference lateral overlap ratio, and then obtain a complete space of the aerial photography area 10. Take a photo.

As shown in FIG. 6, if the terrain of the aerial photography area 30 is judged as flat land, a flat land route planning mode is executed. The flat land route planning mode is used to generate a flat land route path 31, wherein the flat land route path 31 includes a plurality of flat land routes. 311. The number of such flat land routes 311 is calculated by the following method: (S201) Calculate the route distance by using the preset reference lateral overlap ratio B%, where the route distance is the distance between the route and the route. The direction overlap rate B% is a preset value. In a preferred embodiment, the side overlap rate B% is preset to 50%; (S202): using the preset one-way overlap rate A% to calculate the baseline length, In a preferred embodiment, the front-to-back overlap ratio A% is preset to 70%; according to the calculated route distance and the baseline length, the number of flat routes 311 can be obtained, and the flight of the flat route 31 is planned. Path, one aerial camera 32 is flying along the flatland route path 31, the flatland route 31 is flying in a shooting direction and a course direction, and the aerial camera 32 is shooting one in each baseline length in the shooting direction Photos, when the length of a flat ground course 31 is taken, it will return to take a reverse shot along the next flat ground course 311, and sequentially take along the flat ground course 311 in the direction of the route to obtain a photo of the range of the aerial area 30 .

This creation is preset to optimize the front-to-back overlap ratio and lateral overlap ratio, and use the numerical elevation model to obtain the maximum height difference in the aerial photography area, to calculate the number of routes, and to use the contour data of the numerical elevation model to plan a match Aerial heights of different contour heights, while adjusting the aerial height of the aerial camera, can keep the resolution of all the photos taken on the ground the same, improving the clarity and accuracy of the overall drawing. At the same time, using the preset overlap rates, whether it is flat terrain or complex hillside terrain, you can plan an appropriate aerial photography path, so that the aerial camera can not only save time and power, but also provide photos that meet the standards of each overlap rate. Keep superimposed photos with the same quality and resolution.

10‧‧‧Aerial shooting area

11‧‧‧ Hillside route route

111‧‧‧ Hillside route

12‧‧‧ aerial camera

20‧‧‧Aerial shooting area

21‧‧‧First Hillside Route

22‧‧‧ Second Hillside Route

23‧‧‧Second Hillside Correction Course

30‧‧‧Aerial shooting area

31‧‧‧ Flat route route

311‧‧‧flat course

32‧‧‧ aerial camera

71‧‧‧ aerial camera

a, b, c‧‧‧overlap area

d, e, f‧‧‧ laterally overlapping areas

path1‧‧‧first path

path2‧‧‧second path

path3‧‧‧second path

path4‧‧‧second path

path5‧‧‧second path

Di‧‧‧ route spacing

g, i‧‧‧height difference

Figure 1: Flow chart of the steps of the route planning method for aerial photography using a multi-axis unmanned aerial vehicle. Figure 2: Schematic diagram of the route of the hillside in this creation. Figure 3: Schematic diagram of the route of this creative contour line. Figure 4: Schematic diagram of the pitch of this creative route before correction. Figure 5: Schematic diagram of the revised route spacing. Figure 6: Schematic diagram of the flat land route of this creation. Figure 7: Schematic diagram of inconsistent overlap rates caused by the prior art. Figure 8: Schematic diagram of inconsistent lateral overlap caused by the prior art. Figure 9: Schematic diagram of inconsistent resolution caused by the prior art.

Claims (6)

A route planning method for aerial photography using a multi-axis unmanned aerial vehicle includes: (a) determining whether the terrain of an aerial photography area is flat or hillside, and according to a maximum altitude value and a minimum altitude value within the aerial photography area; A first height difference between the two, determines that the terrain of the aerial photography area is flat or hillside; when the aerial photography area is judged to be hillside, a hillside route planning mode is executed to generate a hillside route path of the aerial photography area, The route of the hillside route includes a plurality of hillside routes. The hillside route planning model involves the following steps: (b1) determining a first hillside route as a straight line or a contour line; (b2) calculating the number of such hillside routes, where The distance between each route is the distance between two adjacent hillside routes. When the aerial photography area is judged as flat ground, a flatland route planning mode is executed to generate a flatland route path in the aerial photography area. The flatland route path includes a plurality of flatland routes. The route planning method for aerial photography using a multi-axis unmanned aerial vehicle as described in claim 1, wherein, in step (a), the first height difference value is divided by a set value to obtain a calculated value, and the calculation The value is compared with a preset ground resolution value (GSD). If the calculated value is greater than the ground resolution value, it is judged as a hillside; if the calculated value is less than the ground resolution value, it is judged as a flat land. The route planning method for aerial photography using a multi-axis unmanned aerial vehicle as described in claim 2. In step (b1), the method includes: defining a surface of a boundary of the aerial photography area as a surface profile line, and defining The route path corresponding to the surface profile is the first hillside route, where the boundary is the boundary adjacent to the lowest altitude value in the aerial photography area; assuming that the first hillside route is a straight line, one of the first hillside routes is calculated. And the second height difference is compared with a route threshold value, where the second height difference value is the difference between the highest altitude value and the lowest altitude value in the surface profile, and the route threshold value is determined by a depth axis The accuracy formula is calculated; if the second height difference is less than the threshold of the route, the first hillside route is a straight line; if the second height difference is greater than the threshold of the route, the first hillside route is a contour line. The route planning method for aerial photography using a multi-axis unmanned aerial vehicle as described in claim 3, in step (b2), includes: (b21) After the aerial photography area is projected into a plane, the aerial photography area Divide the distance in the direction of the route by the distance between the routes and add 1 to calculate the number of such hillside routes; (b22) Divide the first height difference by the number of these hillside routes to obtain the distance between two adjacent route paths. Corresponding to the initial elevation value of the contour line, the aerial height of the first hillside route is sequentially added to the initial altitude value, and the aerial heights of the hillside routes are obtained one by one; Whether the overlap rate meets the preset reference lateral overlap rate. If the actual distance between two adjacent hillside routes is greater than the calculated route distance, the aerial height of the next hillside route will be reduced to make the two adjacent hillside routes The actual distance is not greater than the distance between the routes, and step (b21) is repeated. The route planning method for aerial photography using a multi-axis unmanned aerial vehicle as described in any one of claims 1 to 4, wherein the flat-land route planning mode is performed with the following steps: (c1) Calculate the route distance, and Calculate the separation distance between the route and the route by the preset lateral overlap ratio; (c2) Calculate the baseline length, and calculate the length of each baseline by the preset front-to-back overlap ratio, where the baseline length is each photo image The length between the centers. The route planning method for aerial photography using a multi-axis unmanned aerial vehicle according to any one of claim 5, wherein the depth axis accuracy formula is: Where σ _{z is} the preset ground resolution value, D is the threshold value of the route, b is the baseline length, d is the camera focal length, and σ _i is the image measurement accuracy.