Depth Map Generation Using Passive and Active Sensor Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing depth camera technologies face challenges in generating accurate depth maps for environments with single-color or plain-pattern regions, as passive sensors struggle to extract features and active sensors consume high power, leading to difficulties in three-dimensional space modeling.
Innovation Solution
An image generating device that combines passive and active sensors, using a mask forming unit to identify and supplement unidentified regions in the depth map, allowing for the generation of a comprehensive depth map by integrating data from both sensors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Use of energy by moving object
If passive sensors are used to obtain depth information, then power consumption is reduced, but measurement precision deteriorates in featureless environments such as single-color or plain-pattern regions
Solution Approach 1:
The patent divides the depth sensing task into two segments: passive sensing for general areas and active sensing for unidentified regions. The processor segments the depth map into identified and unidentified regions, then selectively applies active sensing only to the unidentified portions, thereby reducing overall power consumption while maintaining accuracy where needed.
Solution Approach 2:
Instead of using active sensors for the entire scene, the patent applies active sensing partially only to unidentified regions. This partial action approach reduces power consumption compared to full active sensing, while still achieving complete depth map coverage by combining results from both passive and active sensing.
2Measurement precision
If active sensors are used to obtain depth information, then measurement precision is improved, but power consumption increases
Solution Approach 1:
The patent segments the sensing approach by using passive sensors for the entire scene first, then applying active sensors only to specific unidentified regions. This segmentation allows high precision where needed while minimizing energy consumption by limiting active sensor usage to necessary areas only.
Solution Approach 2:
The patent employs partial action by using active sensors only for unidentified regions rather than the entire scene. This reduces power consumption compared to continuous active sensing, while maintaining measurement precision for the critical areas that require it.
3Use of energy by moving object
If only passive sensors are used, then power consumption is low, but reliability deteriorates in environments with single-color or plain-pattern regions
Solution Approach 1:
The patent merges the outputs of passive and active sensors to create a complete depth map. The processor combines the depth information from passive sensing with the supplementary depth information from active sensing, achieving reliable depth mapping in featureless environments while maintaining lower overall power consumption than active sensing alone.
Solution Approach 2:
The patent uses partial active sensing action only for unidentified regions to supplement passive sensing. This partial intervention improves reliability in featureless areas without the full power consumption cost of continuous active sensing, achieving a balance between reliability and energy efficiency.
4Measurement precision
If a combination of passive and active sensors is used, then measurement precision is improved across all environments, but device complexity increases
Solution Approach 1:
The patent segments the processing workflow into distinct stages: passive sensing, identification of unidentified regions, selective active sensing, and merging of results. This segmentation of the process, rather than requiring complex simultaneous operation of all sensors, manages device complexity while achieving improved measurement precision through coordinated use of both sensor types.
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 enables effective three-dimensional space modeling by overcoming the limitations of passive sensors in featureless environments and reducing power consumption in active sensors, providing accurate depth information across various spaces.
Implementation Method 1
depth cameras employing a time-of-flight (ToF) scheme are used. The depth camera may obtain depth information corresponding to a distance between an object in a captured space and the depth camera.
Implementation Method 2
obtain a second depth map with respect to the unidentified region by using the active sensor
Data Source
AI summary
An image generating device, a method and a computer-readable recording medium are provided. The image generating device includes a passive sensor, an active sensor, a mask forming unit, a memory storing at least one instruction, and a processor configured to execute the at least one instruction to obtain a first depth map with respect to a first space by using the passive sensor, identify an unidentified region of the first depth map, obtain a second depth map with respect to the unidentified region by using the active sensor and the mask forming unit configured to form a mask based on the unidentified region, and obtain a third depth map with respect to the first space based on the first depth map and the second depth map.