Depth Sensing Correction Under Strong External Light
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Robots operating in environments with strong external light, such as airports or schools, face challenges in accurately sensing depth due to distortion caused by external lighting, leading to incorrect distance measurements and image acquisition issues.
Innovation Solution
A method and device that involve storing and comparing depth information at different time points to identify and filter out distortion in depth images, using a controller to adjust depth values and remove incorrect data, and employing an illuminance sensor to determine when to activate image control processes for the depth or vision camera.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If depth sensing is performed in environments with strong external light, then the robot can operate in diverse spaces such as airports and schools, but the depth sensing accuracy deteriorates due to light reflection distortion
Solution Approach 1:
The system performs preliminary depth sensing at a first time point before the object is affected by strong external light, and stores this reference depth information. When distortion is detected at a second time point, the preliminary depth information is used to correct the distorted measurements, preventing the accuracy deterioration before it occurs.
Solution Approach 2:
The system continuously compares depth information from different time points and provides feedback to identify filtering target regions where distortion has occurred. Based on this feedback, the controller adjusts depth values in affected regions, maintaining measurement precision while operating in diverse environments with varying light conditions.
2Measurement precision
If depth information is continuously sensed and stored, then the accuracy of obstacle detection is improved, but the device complexity and data processing load increase
Solution Approach 1:
Instead of processing all stored depth information, the system extracts only the specific depth information from the first time point that corresponds to the filtering target region identified at the second time point. This selective extraction reduces data processing complexity while maintaining obstacle detection accuracy by focusing only on relevant distorted regions.
Solution Approach 2:
The depth sensing data is segmented by time points and regions. The system divides the continuous depth information into discrete time-point snapshots and further segments the spatial data to identify specific filtering target regions. This segmentation allows efficient comparison and correction without processing the entire data set, reducing overall system complexity.
3Loss of information
If the robot uses depth camera and vision RGB camera for 3D and 2D image information, then comprehensive spatial awareness is achieved, but the sensing system becomes more vulnerable to external light distortion
Solution Approach 1:
The system dynamically adjusts the depth sensing operation based on detected distortion conditions. When external light distortion is detected in the vision RGB camera data, the system activates the depth camera to provide corrected depth information for the affected regions, creating a dynamic compensation mechanism that maintains spatial information completeness while counteracting light distortion effects.
Solution Approach 2:
The depth camera acts as an intermediary system that provides accurate depth information to compensate for distortion detected in the vision RGB camera. When external light causes distortion in the 2D image information, the depth camera's depth data serves as a mediator to correct the spatial interpretation, maintaining comprehensive spatial awareness despite the vulnerability of either system alone.
Data Source
AI summary
The present disclosure relates to a method for sensing the depth of an object by considering external light and a device implementing the same, and a method for sensing the depth of an object by considering external light according to an embodiment of the present disclosure comprises the steps of: storing, in a storage unit, first depth information of an object, which is sensed at a first time point by a depth camera unit of a depth sensing module; storing, in the storage unit, second depth information of the object, which is sensed at a second time point by the depth camera unit; comparing, by a sensing data filtering unit of the depth sensing module, the generated first and second depth information to identify a filtering target region from the second depth information; and adjusting, by a control unit of the depth sensing module, the depth value of the region filtered from the second depth information.