The figure below simulates a scene of a square building shot with a short focal length camera. Due to reasons such as short focal length, shooting distance and angle, the deformation of the building in the scene is high, and it is difficult to reconstruct it simply by using the perspective matching function of 3ds max. Shooting the scene, we use the perspective matching function to cooperate with the camera adjustment to perfectly realize the reconstruction of the scene.
Open 3ds max, display four views, click the third menu item above the perspective view, and then select viewport background and configure viewport background in turn, as shown in the figure below:
Select the image file that will serve as the background,
Set the aspect ratio to match the bitmap.
Set the aspect ratio of the rendering resolution to the aspect ratio of the background image through the rendering settings, or make the width and height of the rendering resolution consistent with the width and height of the background image, and then set the display safe frame.
Press alt+w to display four views, convert perspective view to front view and back to perspective view, press ctrl+shift+z to maximize each view.
Draw a vertical line downward from the origin in the top view. The direction of this line is the y-axis direction of the world coordinates. This line will indicate the orientation in the perspective view and serve as a reference for subsequent operations.
Create a cuboid in the top view whose position should be behind and to the left of the origin in the top view.
Set the perspective view to wireframe overlay mode.
In the perspective view, use rotation (alt+middle button drag) and translation operations (middle button drag) to adjust the view as shown in the figure below. Note that only rotation and translation operations can be used, and other operations, such as zooming the view, cannot be used. The same is true for subsequent operations.
Turn on perspective matching.
First adjust the red shadowless line (x-axis), then adjust the green shadowless line (y-axis, refer to the line originally drawn in the view), and finally adjust the blue shadowless line. If the view does not change at this time, you can adjust it slightly At one end of a blue line, the view will change, and the position and inclination of the cube will change greatly, as shown in the figure below:
Use translation (middle key drag) to adjust the cube in the view to align with the front, bottom, right point of the cube in the background. Although the sides of the cube are not in harmony with the sides of the background cube, it doesn’t matter, so don’t worry about it.
Press alt+w, only the perspective view will be displayed, so that you can adjust the red and green shadow lines more precisely, you will find that the blue shadow lines will automatically tilt, no need to be physical and chemical, just adjust the red and green shadow lines to make them consistent with The edges of the background cube are aligned.
Press ctrl+c to create a camera based on perspective, press alt+w to display all views, select the camera, enter the modification panel, check the automatic vertical tilt correction in the perspective control panel, at this time, the cube in the camera view is not It's slanted again. Use the translation method to align the cube with the front, bottom, and right points of the background cube. Note that after the camera view, you can only use translation (drag with the middle button) to adjust, and you cannot use rotation (alt+drag with the middle button), otherwise it will Messed up.
Convert the cube into an editable polygon, enter the polygon level, and move its top surface (z-axis direction), left side (x-axis direction), and back surface (y-axis direction) to make its size basically match the background cube. Each surface can refer to the line used by the view to indicate the direction of the y-axis.
However, the procedure for displaying that the cube is deformed is too high. It is necessary to lengthen the focal length of the camera, increase the focal length a little, and then translate the view so that the bottom edge of the cube is also parallel to the background cube.
Adjust the positions of the faces again.
But the inclination of the upper two sides is too large, adjust the vertical in the perspective control panel of the camera to 1, and pan the view to align the upper and lower endpoints of the center line with the corresponding endpoints of the background cube.
At this time, the inclinations of the upper and lower groups of sides are basically the same relative to the background cube, but the degree of deformation is still a bit high, and the focal length of the camera needs to be adjusted again. Properly lengthen the focal length, translate, adjust the position of each surface, and finally get a satisfactory alignment, and the camera adjustment is completed.
