1. Determination of aerial altitude
The altitude of UAV tilt photography is the basis of route design. The aerial altitude needs to select the appropriate ground resolution according to the requirements of the mission, then combines with the performance of the tilting camera, according to formula (1), calculate the aerial height of H = f × GSD/ α (1), in units of m; f is the lens focal length, unit.
Mm; α is pixel size in mm; GSD is ground resolution in m.
2. Setting of overlap degree of aerial capture
The low-altitude digital aerial photography specification stipulates that the "directional overlap" shall generally be 60 % ~ 80 %, minimum not less than 53 %; the bypass overlap shall generally be 15 % ~ 60 %, minimum not less than 8 %. In case of tilt photography of the aerial camera drone
, the side direction overlap is obviously insufficient. The recommended value is 66.7 % in accordance with the algorithm theory. It can be distinguished into two cases of building sparse area and building dense area.
2.1 Sparsely constructed areas
Considering its pitch and roll influence, the aerial photo drone is not less than 70 % in the survey area without a high-rise building or a terrain figure with a low altitude difference when it is inclined. In order to obtain the complete image information of a certain area, the drone must fly over the area. If the two buildings are not availble for the aircraft to fly over the area due to the height, so the adhesion of the geometric structure of the building model can be caused no matter how many cameras are increased .
2.2 Construction-intensive areas
The problem of building occlusion in built-up areas is very serious. The insufficient design of route overlap and the failure of aerial photography to fly over the relevant buildings will result in the conglutination of the geometric structure of the building model. In order to improve the quality of image acquisition in built-up areas, the maximum degree of image overlap can be designed as 80% ~ 90%. When the height of high-rise building is more than a quarter of the aerial height, the method of increasing image overlap and cross-flying and increasing redundant observation can be adopted to solve the problem. For example, the famous slope photography of Lujiazui area in Shanghai uses more than 90% overlap to collect images to prevent the buildings from blocking each other. The degree of image overlap is closely related to the amount of image data. The higher the degree of image overlap, the larger the amount of data in the same area and the lower the efficiency of data processing. So in the course of route design should also take into account the balance between the two.
2.3 Area coverage design
"Course coverage beyond the area boundary shall be not less than two baselines. The side coverage is generally not less than 50% of the range ", which is the guarantee of the original code on the aerial area boundary coverage aerial,
but it is obviously not enough in the tilt photography of aerial filming drone
. In theory, it is necessary that the edge of the target area can appear in any position of the image, just like the observation of the characteristic point in the central area of the measuring area. Considering the height difference of the measured area, the width L ~ (1 × tan) θ H2-H3 () L ~ (1 × tan θ H2-H3 ()) of the route extension can be calculated according to formula (2). H₁ is relative altitude; θ is camera tilt angle; H₂ is the height of the photographic datum; H₃ is the lowest point height of the edge of the measured area; L1 is the horizontal distance corresponding to half an image amplitude.