Vehicle ambient environment three-dimensional reconstruction method for night vision

Abstract

The invention discloses a vehicle ambient environment three-dimensional reconstruction method for night vision. According to the technical scheme, a night vision system based on a three-dimensional laser radar sensor is constructed, the night vision system is composed of the vehicle mounted three-dimensional laser radar sensor, a computer, a display and three-dimensional reconstruction software, the three-dimensional reconstruction software is composed of an environmental mesh model generating module and a model color rendering module, the environmental mesh model generating module generates an environment height graph through three-dimensional point clouds, a triangular mesh environmental model is generated by conducting triangular patching on the environment height graph, the model color rendering module puts color on each triangular patch in the triangular mesh environmental model so that a three-dimensional reconstruction image can be generated, and the three-dimensional reconstruction image is displayed by the display. By the adoption of the vehicle ambient environment three-dimensional reconstruction method for the night vision, the three-dimensional rendering type terrain around a vehicle can be generated, the color of the reconstructed image can be closer to the visual sense of the eyes of people, display enhancement is conducted on the vision blind areas of a driver, the driver can have a good vision field even in the dark environment, and the requirements of night driving are met.

Description

technical field [0001] The invention relates to the field of graphics and image processing, in particular to a method for three-dimensional reconstruction of the surrounding environment of a vehicle for night vision. Background technique [0002] According to statistics, more than 80% of road traffic accidents occur at night, and most major traffic accidents and escape cases occur at night. The biggest safety hazard of driving at night is that there are many visual blind spots in low light or no light conditions at night, making it impossible for drivers to fully observe and grasp the direction of the road, and it is difficult to see all threatening obstacles around the vehicle (such as guardrails, boulders, Pedestrians, ditches and forward vehicles, etc.) are all under control, resulting in frequent accidents. Therefore, aiming at the problems existing in the driving of the vehicle at night in dim or dark conditions, developing a corresponding scene enhancement system for ...

AI Technical Summary

Problems solved by technology

The cost of high dynamic cameras is low, and auxiliary lighting is required for night use, and additional contrast enhancement and strong light suppression algorithms are required, so the robustness is not strong; low-light night vision devices have the characteristics of small size, light weight and high work safety. Advantages, but its operating distance and observation effect are greatly affected by the environment, and it is not available in a dark environment; active infrared night vision images have the advantages of relatively clear images and high contrast, but they require active infrared lighting, and there is a risk of characteristic exposure , and it is greatly affected by the environment, afraid of direct sunlight

[0005] Traditional lidar data visualization methods mainly include the following two methods: one is direct representation, which converts lidar data into 3D point cloud display or generates 2D depth map display according to the logical relationship of data. This display method is highly abstract and human It is impossible to understand its meaning directly; the second is the grid map representation, which is to project the data onto a two-dimensional grid map, and assign a color to the grid according to a certain attribute of the grid and display it. The model generated by this method is spatially Discontinuous, and the coloring method does not conform to the human understanding method

[0006] To sum up, the current night vision methods generally have the disadvantages of low resolution, low contrast, and unclear edges. Everything a driver needs to drive at night

Images