A method for calculating parameters of unmanned aerial photography used at traffic accident scenes
By calculating the inclination angle and shooting interval angle of the drone, the core area size and drone flight altitude of the traffic accident site are determined, and the problems of traditional low exploration efficiency and low accuracy are solved, and efficient and accurate traffic accident site are achieved. Three-dimensional real-life modeling and exploration.
Patent Information
- Application Number
- CN202410546062.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-04-30
- Publication Date
- 2024-11-05
- Estimated Expiration
- 2044-04-30
AI Technical Summary
Traditional traffic accident site surveys rely on manual photography and measurement, with low efficiency and low accuracy, and cannot achieve secondary re-survey and re-testing, which affects the analysis of the cause of the accident and the determination of responsibility. In addition, there is a lack of quantitative methods for calculation of aerial photography parameters of drones, resulting in image quality It is not high and has low efficiency, and cannot meet the accuracy requirements of three-dimensional real scene modeling.
By calculating the inclination angle β and shooting interval angle δ of the drone gimbal, we determine the core area size of the traffic accident site and the flight altitude and shooting parameters of the drone, and use the envelope hemisphere model to cover the accident site to ensure the image data Integrity and accuracy to reduce flight energy loss.
It realizes the rapid and efficient calculation of drone aerial photography parameters, reduces shooting blind spots and missing detailed information, improves the accuracy of the three-dimensional real scene model and the efficiency of accident site investigation, and adapts to the fast and accurate needs of traffic accident sites.
Smart Images

Figure CN118484024B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of traffic accident scene investigation, and in particular to a method for calculating parameters of unmanned aerial vehicle aerial photography used at traffic accident scenes. Background Art
[0002] Traditional road traffic accident scene investigation mainly relies on manual photography and measurement to fix and extract trace evidence at the scene, which not only occupies road traffic resources, but also easily causes secondary accidents and threatens the personal safety of investigators. Manual measurement and recording have problems such as low efficiency and low accuracy. Once the accident scene is cleared, it is impossible to conduct a secondary re-survey and supplementary measurement, which affects the analysis of the cause of the accident and the determination of responsibility. The application of new accident scene investigation methods to efficiently, comprehensively and objectively record the traffic accident scene has high application value.
[0003] With the deep integration and application of GPS, GIS and drone technology, the use of drone oblique photography technology can realize the three-dimensional real scene modeling of road traffic accident scenes, which can not only break the limitations of two-dimensional image visual presentation, but also complete data measurement based on the constructed model. This technology can effectively solve the problems of manual close-up photography and measurement in the current accident scene investigation work, realize the accident scene re-investigation, and has a high application space in road traffic accident scene investigation.
[0004] Due to the small scope of the road traffic accident scene, the large amount of road horizontal information, the small vertical space, and the great safety hazards, it is required that drone aerial photography not only quickly complete the image acquisition of the traffic accident scene, but also that the acquired image data meet the needs of refined modeling. The traditional drone flight route with a fixed camera angle will have large photography blind spots and missing details due to the complex appearance and structure of the shooting object and mutual occlusion between each other. It requires multiple aerial photography at different angles and heights to collect image data, which is inefficient and poses a great challenge to the drone's endurance. It is not highly compatible with the requirements of timely and rapid investigation of traffic accident sites.
[0005] UAV aerial images are the data basis for 3D real-scene modeling. UAV aerial photography parameters directly determine the quality of UAV aerial photography efficiency image data, and even affect the final modeling effect and accuracy. Currently, UAV points of interest (POI) are used to surround the scene of traffic accidents. Aerial photography parameters are mainly set and adjusted based on manual experience and subjective vision. There is no quantitative calculation method for UAV aerial photography parameters, resulting in low image quality, large number of photos, and low efficiency when obtaining UAV aerial photography of traffic accident scenes. Ultimately, the accuracy of the constructed 3D real-scene model does not meet the measurement error requirements of trace evidence investigation at the scene of the traffic accident. Summary of the invention
[0006] The main purpose of the present invention is to provide a method for calculating parameters of unmanned aerial vehicle aerial photography at a traffic accident scene, which can quickly calculate the parameters required for aerial photography.
[0007] To achieve the above purpose, the technical solution adopted by the present invention is: a method for calculating parameters of drone aerial photography at a traffic accident scene, characterized in that it specifically includes the following steps:
[0008] Step 1: Calculate the gimbal inclination angle β of the gimbal on the drone, including the following steps:
[0009] Step 11, determine the size of the core area of the traffic accident scene, l×w×h, where the core area is a rectangular parallelepiped, l is the length, w is the width, the length and width are the maximum dimensions of the traffic accident scene in two mutually perpendicular directions in the horizontal plane, and h is the height, which is the highest dimension of the traffic accident scene. l, w, and h are obtained by manual measurement, and the core area is determined by the staff according to the on-site conditions;
[0010] Step 12: Take the center position of the bottom surface of the core area as the center point, and construct an envelope hemisphere with the center point as the center of the sphere. The end surface of the envelope hemisphere is located on the bottom surface of the core area. The area covered by the envelope hemisphere is the area that the drone flight needs to shoot. The radius r of the hemisphere is:
[0011]
[0012] Step 13: For the determined road traffic accident scene and the selected UAV model parameters, the minimum flight altitude H is obtained by formula (2): min The relationship between the gimbal tilt angle β is:
[0013] H min =R min sinβ (2),
[0014] Among them, R min is the minimum distance between the drone and the center point when the gimbal inclination angle is β,
[0015] R min =r / sin(FOV / 2) (3),
[0016] FOV is the viewing angle of the drone lens when shooting. When the accident center is taken as the POI point for surround shooting, the center point is on the main optical axis of the drone lens, that is, on the bisector of the FOV;
[0017] Step 14: Calculate the maximum value of the gimbal inclination angle β max :
[0018]
[0019] Wherein, R is the distance between the UAV and the center point;
[0020] Step 15: Substitute formula (3) into formula (4) to obtain β max The minimum value (β max ) min :
[0021]
[0022] Step 16: Get the ground resolution of the center point, near end point, and far end point:
[0023]
[0024]
[0025]
[0026] GSD 中 GSD is the resolution of the captured center point in the image. 远 GSD is the resolution of the far end point in the image. The far end point is the point farthest from the drone within the shooting range. 近 is the resolution of the near end point in the image. The near end point is the point closest to the drone within the shooting range. From formula (7), it can be obtained that β>FOV / 2;
[0027] Step 17, substitute formula (2) into formula (7) to obtain:
[0028]
[0029] Step 18: For a single image, the ground resolution of the far point is no more than five times that of the center point, which can be specifically expressed as follows:
[0030] GSD 远 ≤5GSD 中 (10);
[0031] Step 19: Substitute formula (2), formula (6), and formula (7) into formula (10) to obtain:
[0032]
[0033] Step 110: Through steps 16, 18 and 19, we can obtain:
[0034]
[0035] Among them, min() means taking the minimum value;
[0036] Step 111: Obtain the difference GSD between the far-end and near-end resolutions in a single image Δ , which can be specifically expressed as follows:
[0037] GSD Δ =GSD 远 -GSD 近 (13);
[0038] Step 112: Substitute formula (7) and formula (8) into formula (13), and according to the value range of β in formula (13), obtain GSD Δ The beta value at the minimum.
[0039] Preferably, the step 1 further includes step 113, substituting the determined β into formula (2) to obtain H min .
[0040] Preferably, the method further includes step 2, calculating the shooting interval angle δ of the drone, and taking a photo once when the drone rotates δ, which specifically includes the following steps:
[0041] Step 21: Get the overlap between two adjacent images:
[0042]
[0043] Among them, ψ is the image overlap, l1 is the long side of the image; l2 is the short side of the image, l1 and l2 are determined by the camera parameters;
[0044] Step 22: Calculate δ according to the required image overlap.
[0045] Compared with the prior art, the present invention has the following beneficial effects:
[0046] The method of the present invention can quickly calculate the parameters of the UAV's circling flight, and reduce the energy loss of the UAV during the flight as much as possible while ensuring that the accident scene can be photographed completely. BRIEF DESCRIPTION OF THE DRAWINGS
[0047] Figure 1 This is a schematic diagram of an aerial photo of a traffic accident scene;
[0048] Figure 2 It is a cross-sectional view of the traffic accident scene taken from an aerial photograph;
[0049] Figure 3 It is an aerial cross-section diagram at different gimbal inclination angles;
[0050] Figure 4 and Figure 5 It is a schematic diagram of the overlap of two different interval angles;
[0051] Figure 6This is the schematic diagram for determining aerial photography parameters. DETAILED DESCRIPTION
[0052] The following description is used to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments described below are only examples, and those skilled in the art may think of other obvious variations.
[0053] A modeling method for calculating a parameter model of a drone aerial photography of a traffic accident scene specifically comprises the following steps:
[0054] Step 1: Calculate the gimbal inclination angle β of the gimbal on the drone, and then adjust the shooting angle of the camera set on the gimbal, 0°<β<90°, specifically including the following steps:
[0055] Step 11, determine the size of the core area of the traffic accident scene, l×w×h, where the core area is a rectangular parallelepiped, l is the length, w is the width, the length and width are the maximum dimensions of the traffic accident scene in two mutually perpendicular directions in the horizontal plane, and h is the height, which is the highest dimension of the traffic accident scene. l, w, and h can be obtained by manual measurement, and the core area can be determined by the staff according to the on-site conditions.
[0056] Step 12: Take the center position of the bottom surface of the core area as the center point, and construct an envelope hemisphere with the center point as the center of the sphere. The end surface of the envelope hemisphere is located on the bottom surface of the core area. The area covered by the envelope hemisphere is the area that the drone flight needs to shoot. The radius r of the hemisphere is:
[0057]
[0058] Step 13: For the determined road traffic accident scene and the selected UAV model parameters, in order to make the aerial image cover the entire core area of the accident scene and reduce the flight altitude of the UAV as much as possible, the minimum flight altitude H can be obtained by formula (2): min The relationship between the gimbal tilt angle β is:
[0059] H min =R min sinβ (2),
[0060] Among them, R min is the minimum distance between the drone and the center point when the gimbal inclination angle is β,
[0061] R min =r / sin(FOV / 2) (3),
[0062] FOV is the viewing angle of the drone lens when shooting. When the accident center is taken as the POI point for surround shooting, the center point is on the principal optical axis of the drone lens, that is, on the bisector of the FOV. When the tilt angle of the gimbal is β, and the minimum distance between the drone and the center point is R min When , the far side edge of FOV is tangent to the spherical surface of the envelope hemisphere, that is, the captured image should at least cover the area covered by the envelope hemisphere;
[0063] Step 14: To avoid the vertical obstruction of the core area of the accident scene during drone aerial photography and reduce the blind spot, the maximum value of the gimbal inclination angle β can be calculated. max for:
[0064]
[0065] Wherein, R is the distance between the drone and the center point. When R is determined, when β takes the maximum value, the edge of the near side of the FOV is just tangent to the envelope hemisphere and perpendicular to the bottom surface. If β continues to increase, a shooting blind spot will be generated.
[0066] Step 15: Substitute formula (3) into formula (4) to obtain β max The minimum value (β max ) min :
[0067]
[0068] Step 16: Get the ground resolution of the center point, near end point, and far end point:
[0069]
[0070]
[0071]
[0072] GSD 中 GSD is the resolution of the captured center point in the image. 远 GSD is the resolution of the far end point in the image. The far end point is the point farthest from the drone within the shooting range. 近 is the resolution of the far end point in the image, and the near end point is the point closest to the drone within the shooting range. Since the resolution value cannot be negative or zero, it can be seen from formula (7) that only when β>FOV / 2, GSD 远 Only then will it have meaning;
[0073] Step 17, substitute formula (2) into formula (7) to obtain:
[0074]
[0075] From formula (9), we can see that as β increases, GSD 远 Gradually decrease, that is, the smaller β is, the better;
[0076] Step 18: In order to ensure the accuracy of the later measurement of the constructed 3D real scene model, for a single image, the ground resolution of the far point should not exceed five times that of the center point, which can be specifically expressed as follows:
[0077] GSD 远 ≤5GSD 中 (10)
[0078] Step 19: Substitute formula (2), formula (6), and formula (7) into formula (10) to obtain:
[0079]
[0080] Step 110: Through steps 16, 18 and 19, we can obtain:
[0081]
[0082] Among them, min() means taking the minimum value;
[0083] Step 111: for the difference GSD between the far end and the near end in a single image Δ It should not be too large, and can be specifically expressed as follows:
[0084] GSD Δ =GSD 远 -GSD 近 (13);
[0085] Step 112: Substitute formula (7) and formula (8) into formula (13), and according to the value range of β in formula (13), obtain GSD Δ The value of β at the minimum;
[0086] Step 113: Substitute the determined β into formula (2) to obtain H min , that is, the cruising altitude of the UAV is H min , the patrol radius is R min , the gimbal inclination angle is β.
[0087] Step 2: Calculate the shooting interval angle δ. The drone takes a shot every time it rotates δ. The smaller δ is, the higher the shooting frequency is, and the more comprehensive the images are. However, the amount of data will increase, and it will take longer to complete all the shooting. The larger δ is, the lower the shooting frequency is, and the fewer images are obtained. However, the data is relatively small, and it will take less time to complete all the shooting. In practice, it is necessary to ensure that the data is comprehensive and the time is shortened as much as possible:
[0088] The overlap between two adjacent images is obtained by the following formula:
[0089]
[0090] Where ψ is the image overlap, l1 is the long side of the image, l2 is the short side of the image, and l1 and l2 are determined by the camera parameters;
[0091] Step 23: Calculate δ according to the required image overlap. In practice, a default image overlap can be set or the image overlap can be adjusted.
[0092] The derivation process of formula (17) is as follows:
[0093] For aerial photography around POIs, the ratio of the overlapping area of two images taken at adjacent points at an angle of δ to the image area is directly reflected in the overlap degree ψ of oblique photography, that is, the ratio of the overlapping area to the image area, as shown in the following formula:
[0094]
[0095] Where: S is the overlapping area of adjacent images, is the image area size;
[0096]
[0097] when When, see Figure 4 :
[0098] The size of the overlapping area S is:
[0099]
[0100] Based on the geometric relationship of rotation, we know that:
[0101] ∠CAB=δ
[0102] ∴AE=AE'=OE·tan(δ / 2)=l1 / 2·tan(δ / 2)
[0103] ∴AB=BE'-AE'=l2-l1 / 2·tan(δ / 2)
[0104]
[0105] Similarly, we can obtain:
[0106]
[0107] So we can get:
[0108] S=l2·l1-[l1 / 2-l2 / 2·tan(δ / 2)] 2 ·tanδ-[l2 / 2-l1 / 2·tan(δ / 2)] 2 tanδ
[0109] It is known that hour:
[0110]
[0111] Simplifying, we get:
[0112]
[0113] when When, see Figure 5 :
[0114] The size of the overlapping area S is:
[0115]
[0116]
[0117] ZP=l2
[0118] Based on the geometric relationship of rotation, we know that:
[0119] ∠OMT=δ
[0120] ∴
[0121] ∴
[0122] in:
[0123]
[0124] ∴
[0125] So we can get:
[0126]
[0127] It is known that hour:
[0128]
[0129] After determining the circling height, circling radius, gimbal inclination angle and interval angle, place the drone at the center point to enable the drone to automatically take pictures of the accident scene.
[0130] In practice, since the parameters of the drone's camera are fixed, it is only necessary to measure the size of the accident scene each time, input the measured size into the drone or remote controller, and then calculate the corresponding cruise parameters.
[0131] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions only describe the principles of the present invention. The present invention may be subject to various changes and improvements without departing from the spirit and scope of the present invention. These changes and improvements fall within the scope of the present invention. The scope of protection claimed by the present invention is defined by the attached claims and their equivalents.
Claims
1. A method for calculating parameters of unmanned aerial photography of traffic accident scenes, characterized in that: The specific steps include: Step 1: Calculate the gimbal inclination angle β of the gimbal on the drone, including the following steps: Step 11, determine the size of the core area of the traffic accident scene, l×w×h, where the core area is a rectangular parallelepiped, l is the length, w is the width, the length and width are the maximum dimensions of the traffic accident scene in two mutually perpendicular directions in the horizontal plane, and h is the height, which is the highest dimension of the traffic accident scene. l, w, and h are obtained by manual measurement, and the core area is determined by the staff according to the on-site conditions; Step 12: Take the center position of the bottom surface of the core area as the center point, and construct an envelope hemisphere with the center point as the center of the sphere. The end surface of the envelope hemisphere is located on the bottom surface of the core area. The area covered by the envelope hemisphere is the area that the drone flight needs to shoot. The radius r of the hemisphere is: Step 13: For the determined road traffic accident scene and the selected UAV model parameters, the minimum flight altitude H is obtained by formula (2): min The relationship between the gimbal tilt angle β is: H min =R min ·sinβ (2), Among them, R min is the minimum distance between the UAV and the center point when the gimbal inclination angle is β, R min =r / sin(FOV / 2) (3), FOV is the viewing angle of the drone lens when shooting. When the accident center is used as the POI point for surround shooting, the center point is on the principal optical axis of the drone lens, that is, on the bisector of the FOV.
2. The method for calculating parameters of unmanned aerial vehicle aerial photography at a traffic accident scene according to claim 1, characterized in that: The step 1 also includes step 14, calculating the maximum value β of the tilt angle of the gimbal max : Wherein, R is the distance between the UAV and the center point; Step 15: Substitute formula (3) into formula (4) to obtain β max The minimum value (β max ) min : Step 16: Get the ground resolution of the center point, near end point, and far end point: GSD 中 GSD is the resolution of the captured center point in the image. 远 GSD is the resolution of the far end point in the image. The far end point is the point farthest from the drone within the shooting range. 近 is the resolution of the near end point in the image. The near end point is the point closest to the drone within the shooting range. From formula (7), it can be obtained that β>FOV / 2; Step 17, substitute formula (2) into formula (7) to obtain: Step 18: For a single image, the ground resolution of the far point is no more than five times that of the center point, which can be specifically expressed as follows: GSD 远 ≤5GSD 中 (10); Step 19: Substitute formula (2), formula (6), and formula (7) into formula (10) to obtain: Step 110: Through steps 16, 18 and 19, we can obtain: Among them, min() means taking the minimum value.
3. The method for calculating parameters of unmanned aerial vehicle aerial photography for traffic accident scenes according to claim 2, characterized in that: Step 111: Obtain the resolution difference GSD between the far end point and the near end point in a single image Δ , which can be specifically expressed as follows: GSD Δ =GSD 远 -GSD 近 (13); Step 112: Substitute formula (7) and formula (8) into formula (13), and according to the value range of β in formula (13), obtain GSD Δ The beta value at the minimum.
4. The method for calculating parameters of unmanned aerial vehicle aerial photography for traffic accident scenes according to claim 3, characterized in that: The step 1 also includes step 113, substituting the determined β into formula (2) to obtain H min .
5. The method for calculating parameters of unmanned aerial vehicle aerial photography for traffic accident scenes according to claim 4, characterized in that: The second step is to calculate the shooting interval angle δ of the drone. The drone takes a shot every time it rotates δ. The specific steps are as follows: Step 21: Get the overlap between two adjacent images: Among them, ψ is the image overlap, l1 is the long side of the image; l2 is the short side of the image, l1 and l2 are determined by the camera parameters; Step 22: Calculate δ according to the required image overlap.
