Method for acquiring distinct scene graph by fusing accident close-range graph with aerial image

Abstract

The invention provides a method for acquiring a distinct scene graph by fusing an accident close-range graph with an aerial image. The method comprises the following steps of S1, placing a marking plate at a considered trace position of an accident scene and acquiring aerial image contents containing global image information of the accident scene by using an aerial technology; S2, carrying out close-range photographing on the accident scene where the marking plate is placed at the considered trace position to acquire a distinct close-range image of the considered trace accident scene and transforming the aerial image and the close-range image to an equal-proportion live-action coordinate system by using homogeneous coordinate transformation; S3, searching and positioning in an aerial image top view and a close-range image top view which are subjected to image transformation by using an image recognition algorithm and acquiring the position of an index marking plate according to the requirement; S4, taking the marking plate as a position index positioning object, and meanwhile, rotating and zooming the close-range image top view subjected to the image transformation to realize the matching of the close-range image top view with a corresponding area of the aerial image in size and direction.

Description

Technical field [0001] The invention relates to the technical field of information acquisition on the scene of a traffic accident, in particular to a method for obtaining a clear scene map by fusing a close-up view of an accident with an aerial image. Background technique [0002] Aerial images play an important role in the survey of traffic accident scenes. The two-dimensional equal-scale real scene map obtained by geometric correction and coordinate transformation can effectively restore the various elements of the accident scene. However, in order to obtain global information, aerial images are often taken at a higher location. , So the ability to distinguish small traces on the ground is insufficient. In the current traffic accident site survey, it is often necessary to photograph the details of each element of the accident. However, due to the random shooting angle and framing range, the location of the captured elements in the entire scene can only be roughly estimated by m...

AI Technical Summary

Problems solved by technology

[0002] Aerial images play an important role in traffic accident scene surveys. The two-dimensional equal-scale real scene map obtained through geometric correction and coordinate transformation can efficiently restore the elements of the accident scene. However, in order to obtain global information, the aerial images are often shot at a higher location. , so the ability to distinguish small traces on the ground is insufficient

In the current traffic accident scene survey, it is often necessary to photograph the details of the various elements of the accident. However, due to the arbitrary shooting angle and viewing range, the position of the photographed elements in the entire scene can only be roughly estimated by manual identification, which is not conducive to Follow-up accident analysis

Images