Monocular Depth Correction Using Road Surface Distance Estimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The depth information generated by a monocular camera is inferior in accuracy of the absolute distance, as observed in existing technologies.
Innovation Solution
An information processing device that includes a depth information generation unit, a road surface distance estimation unit, and a depth information correction unit, which generates depth information based on a captured image from a monocular camera, estimates the road surface distance, and corrects the depth information using the estimated road surface distance.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If depth information is generated using a monocular camera, then the device complexity is reduced compared to stereo cameras, but the measurement precision of absolute distance deteriorates
Solution Approach 1:
The patent introduces road surface distance as an intermediary parameter to bridge the gap between monocular camera imaging and actual three-dimensional space. By estimating the distance to the road surface and using it as a reference, the system can correct depth information from the monocular camera to achieve accurate absolute distance measurement without requiring complex stereo camera systems.
Solution Approach 2:
The patent changes the parameter space by introducing road surface distance estimation as an additional parameter. Instead of directly relying on monocular depth information which lacks absolute scale, the system estimates road surface distance and uses it to transform and correct the depth information, thereby achieving accurate absolute distance measurement with a simple monocular camera.
2Measurement precision
If stereo camera is used to generate accurate depth information, then the measurement precision of depth information is improved, but the device complexity and cost increase
Solution Approach 1:
The patent extracts the essential function of distance measurement from the complex stereo camera system and implements it separately through road surface distance estimation. By taking out the road surface distance estimation function and using it as a reference, the system achieves accurate depth measurement without requiring the complex stereo camera hardware.
Solution Approach 2:
The patent creates a virtual reference frame by estimating road surface distance and using it to correct monocular depth information. This copying approach allows the system to achieve stereo-camera-level accuracy by creating a reference model of the road surface geometry, which can then be used to transform monocular images into accurate three-dimensional coordinates.
Data Source
AI summary
The present technology relates to an information processing device, an information processing method, and an information processing system that make it possible to improve the accuracy of depth information generated by use of a monocular camera. The information processing device includes a depth information generation unit that generates depth information on the basis of a captured image captured by a monocular camera provided on a vehicle such that the camera is able to capture an image of a road surface, a road surface distance estimation unit that executes an estimation process of a road surface distance that is a distance to the road surface in the captured image, and a depth information correction unit that corrects the depth information on the basis of a result of the estimation of the road surface distance. The present technology can be applied, for example, to a vehicle.