Depth Map Fusion Using Ambient Light Indication
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing depth sensors often produce suboptimal and inaccurate depth information, leading to artefacts in three-dimensional image rendering and processing, particularly in autostereoscopic displays, due to their suboptimal performance and the need for dedicated sensors, which increases cost and reduces flexibility.
Innovation Solution
Combining active and passive depth sensors using an ambient infrared light indication to generate improved depth maps, allowing for a weighted combination of depth values from both sensors, which can be achieved with standard off-the-shelf equipment without requiring additional infrared measurement functionality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If dedicated modified depth sensors are used to improve depth estimation accuracy, then measurement precision improves, but device complexity and cost increase
Solution Approach 1:
The patent combines data from multiple standard depth sensors (e.g., stereo camera, TOF sensor, structured light sensor) to generate a unified depth map. By merging information from different sensing modalities, the system achieves improved depth accuracy without requiring dedicated modified sensors, thus resolving the contradiction between measurement precision and device complexity
Solution Approach 2:
The system uses standard off-the-shelf depth sensors that can be applied across multiple applications and scenarios. These universal sensors perform multiple functions (depth mapping, scene understanding, etc.) and their data is processed through a unified framework that adapts to different lighting conditions and environments, achieving high accuracy without specialized hardware
2Measurement precision
If multiple depth sensors are combined to improve depth map quality, then measurement precision improves, but device complexity increases
Solution Approach 1:
The patent introduces an intermediary processing framework that receives data from multiple depth sensors and generates a unified depth map. This intermediary system handles the complexity of sensor fusion, data reconciliation, and quality assessment, allowing standard sensors to be combined effectively without directly increasing system complexity at the hardware level
Solution Approach 2:
The system dynamically adjusts processing parameters (such as weighting factors, fusion methods, and quality thresholds) based on lighting conditions and sensor performance characteristics. By changing parameters rather than hardware configuration, the system adapts to different scenarios while maintaining simplicity in the core architecture
3Device complexity
If standard off-the-shelf sensors are used to reduce cost and complexity, then device complexity decreases, but measurement precision worsens
Solution Approach 1:
The patent combines data from multiple standard depth sensors (e.g., stereo camera, TOF sensor, structured light sensor) to generate a unified depth map. By merging information from different sensing modalities, the system achieves improved depth accuracy without requiring dedicated modified sensors, thus resolving the contradiction between measurement precision and device complexity
Solution Approach 2:
The system creates a composite depth map by integrating data from multiple sensor types with different strengths and weaknesses. This composite approach leverages the complementary characteristics of various standard sensors (passive stereo for texture-rich areas, active TOF for distance measurement, structured light for geometric accuracy) to achieve high overall precision while using only commercially available components
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach provides more reliable and accurate depth values, enhancing the quality of three-dimensional image rendering and processing, improving the perceived image quality and flexibility, while reducing complexity and cost by utilizing standard sensors.
Implementation Method 1
an active depth sensor which includes infrared light emitters which project an infrared light pattern on the scene being recorded
Implementation Method 2
an active depth sensor may comprise a light emitter emitting infrared light in different directions. The time of arrival for reflected light in the different directions may be detected and used to derive depth information
Implementation Method 3
Specific depth values may then be generated by estimating disparities between corresponding image objects in the two images
Data Source
Figure 1
Figure 2
AI summary
An apparatus for determining a depth map for an image of a scene comprises an active depth sensor (103) and a passive depth sensor (105) for determining a depth map for the image. The apparatus further comprises a light determiner (109) which determines a light indication indicative of a light characteristic for the scene. The light indication may specifically reflect a level of visible and/or infrared light for the scene. A depth map processor (107) determines an output depth map for the image by combining the first depth map and the second depth map. Specifically, a depth value for the output depth map is determined as a combination of depth values of the first and the second depth map where the combining is dependent on the light indication. The light determiner (109) estimates the first infrared light indication from a visible light indication indicative of a light level in a frequency band of the visible light spectrum.