Surround View Bowl Mapping for Above-Ground Distortion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Surround view systems in automotive applications suffer from significant distortion when rendering objects above the ground plane, such as vehicles, pedestrians, and signs, due to the use of a virtual bowl surface that is not accurately aligned with these objects, leading to the 'monster truck effect', and require high computational power for regenerating bowl to physical camera mappings.
Innovation Solution
A digital signal processor (DSP) generates the bowl to physical camera mapping, reducing the load on the graphics processing unit (GPU) by 5 to 8 times, allowing for improved performance, lower power consumption, and reduced GPU size, while a processor updates the virtual bowl surface based on depth measurements to mitigate distortion.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If the virtual bowl surface is used for rendering surround view, then the ground plane objects are rendered correctly, but objects above the ground plane suffer from significant distortion (monster truck effect)
Solution Approach 1:
The virtual bowl surface is dynamically adjusted based on detected objects. When objects above the ground plane are detected, the system modifies the bowl surface parameters (depth, curvature) to better align with the actual object positions, transforming the static rendering surface into a dynamic one that adapts to scene content.
Solution Approach 2:
The system changes the parameters of the virtual bowl surface (such as depth values, curvature radii, and surface height) to match the detected object positions. By adjusting these geometric parameters, the rendering surface better conforms to the actual spatial distribution of objects, reducing distortion for objects above the ground plane.
2Manufacturing precision
If the bowl to physical camera mapping is regenerated frequently to improve rendering accuracy, then the distortion is reduced, but the computational power requirement increases significantly
Solution Approach 1:
The system implements feedback by continuously monitoring object positions and depths in the scene. Based on this feedback, it selectively updates the virtual bowl surface and recalculates mappings only when necessary (when objects above ground plane are detected), rather than regenerating mappings for every frame, thus reducing computational load while maintaining accuracy.
Solution Approach 2:
The system performs preliminary detection of objects above the ground plane and pre-adjusts the virtual bowl surface parameters before rendering. This preliminary action prevents the need for frequent full recalculations during rendering, as the mapping is already optimized based on detected objects.
3Productivity
If multiple processors are used to handle depth measurements and rendering, then the processing speed increases, but the device complexity increases
Solution Approach 1:
The system introduces a virtual bowl surface as an intermediary data structure that mediates between depth measurements from sensors and the final rendered image. This intermediary allows different processing stages (depth processing, mapping generation, rendering) to work independently and efficiently, improving overall productivity while managing complexity through clear separation of concerns.
Data Source
AI summary
A system on a chip (SoC) includes a digital signal processor (DSP) and a graphics processing unit (GPU) coupled to the DSP. The DSP is configured to receive a stream of received depth measurements and generate a virtual bowl surface based on the stream of received depth measurements. The DSP is also configured to generate a bowl to physical camera mapping based on the virtual bowl surface. The GPU is configured to receive a first texture and receive a second texture. The GPU is also configured to perform physical camera to virtual camera transformation on the first texture and on the second texture, based on the bowl to physical camera mapping, to generate an output image.