Onboard Camera Attitude Estimation Using Road-Shaped ROI
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional techniques for estimating the attitude of onboard cameras using optical flows from feature points on road surfaces are prone to inaccuracies due to the extraction of false flows, especially when rectangular regions of interest are used, leading to misalignment in rotation angles and reduced estimation accuracy.
Innovation Solution
The proposed solution involves performing initial attitude estimation using a rectangular region of interest and subsequently switching to a superimposed region of interest aligned with the road surface shape, reducing false flows and improving estimation accuracy by using optical flows from the superimposed region.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If a rectangular region of interest is used for attitude estimation, then the processing is simple and straightforward, but false optical flows are extracted leading to reduced estimation accuracy
Solution Approach 1:
The region of interest is segmented into multiple sub-regions based on road surface detection. Instead of processing the entire rectangular ROI, the system divides it into valid road surface areas and invalid areas, processing only the valid portions to extract feature points and optical flows, thereby eliminating false flows from non-road areas.
Solution Approach 2:
Different regions within the ROI are treated differently based on their content. Road surface regions are designated as valid for attitude estimation while non-road regions (sky, buildings, etc.) are marked as invalid. This local differentiation ensures that optical flows are only extracted from appropriate areas, improving accuracy without excessive complexity.
2Measurement precision
If a region of interest aligned with road surface shape is used, then false optical flows are reduced improving estimation accuracy, but the processing complexity increases
Solution Approach 1:
The system performs preliminary road surface detection and validation before extracting feature points and optical flows. By pre-identifying valid road surface regions and marking invalid areas, the system prepares the ROI in advance, so that subsequent optical flow extraction only processes valid regions, reducing false flows without requiring complex real-time adjustments.
Solution Approach 2:
The system automatically detects and validates road surface regions within the ROI, eliminating the need for manual configuration or complex external processing. The self-validation mechanism identifies valid vs. invalid regions and adjusts the processing accordingly, achieving high accuracy through automated, relatively simple operations.
3Productivity
If optical flows are extracted from the entire rectangular region, then processing is fast and efficient, but misalignment in rotation angles occurs reducing reliability
Solution Approach 1:
The system extracts and removes invalid regions (non-road surfaces) from the ROI, isolating only the valid road surface areas for optical flow extraction. By taking out the problematic non-road regions, the system prevents false optical flows from being generated, ensuring reliable rotation angle alignment while maintaining processing efficiency through focused computation on valid areas only.
Data Source
AI summary
An information processing apparatus according to the embodiment includes a control unit (corresponding to an example of a “controller”). The control unit performs attitude estimation processing to estimate the attitude of an onboard camera based on optical flows of feature points in a region of interest set in an image captured by the onboard camera. When the onboard camera is mounted in a first state, the control unit performs first attitude estimation processing using a first region of interest set in a rectangular shape, and, when the onboard camera is mounted in a second state, the control unit performs second attitude estimation processing using a second region of interest set in accordance with the shape of a road surface.