Depth Map Generation Using Time Window Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing depth map generation methods using time of flight (ToF) face challenges in achieving high precision while maintaining a high frame rate, as they require large memory capacity and multiple measurements per frame, which can reduce measurement speed and increase memory constraints, especially with a large number of sensors.
Innovation Solution
A depth map generation apparatus and method that divides the acquisition time of sensors into multiple time windows, allowing for coarse and fine measurements to determine object detection, thereby generating sub-frame information efficiently and reducing unnecessary measurements, and transmitting this information as a single bit data packet to improve processing speed and memory efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple measurements are performed to increase depth map precision, then measurement precision is improved, but frame rate is reduced
Solution Approach 1:
The patent divides the sensor array into multiple regions of interest (ROIs), each processed by a dedicated control circuit. This segmentation allows parallel processing of different spatial regions, enabling multiple measurements to be performed simultaneously across different ROIs, thereby maintaining high frame rate while achieving precise depth mapping through multiple measurements per frame.
2Area of stationary object
If a large number of sensors are used to increase measurement area, then area of measurement is improved, but memory capacity is increased
Solution Approach 1:
The patent divides the large sensor array into multiple smaller regions of interest (ROIs), with each ROI processed independently by a dedicated control circuit. This segmentation reduces the memory burden on any single processing unit, as each control circuit only needs to store and process data for its assigned ROI rather than the entire sensor array, enabling large-scale sensor arrays to be managed with limited memory resources.
Solution Approach 2:
The patent introduces a spatial dimension by dividing the measurement area into multiple regions of interest (ROIs) that can be processed in parallel. This dimensional organization transforms the memory problem from a single large buffer requirement into multiple smaller buffers, effectively distributing memory usage across spatial segments and reducing the peak memory capacity required.
3Measurement precision
If high capacity memory is used to improve depth map precision, then measurement precision is improved, but device complexity is increased
Solution Approach 1:
The patent divides the sensor array into multiple regions of interest (ROIs), each handled by a dedicated control circuit with its own smaller memory buffer. This segmentation replaces the need for a single large high-capacity memory with multiple smaller memory units, achieving the same precision goal while reducing overall device complexity and memory cost.
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 enhances depth map generation speed and precision by optimizing memory usage and reducing data transmission time, allowing for faster and more accurate depth information acquisition without the need for extensive memory resources.
Implementation Method 1
uses, to obtain depth information, the difference, or 'ToF ', between the light irradiation time and the arrival time of reflected light after irradiating light to an object
Implementation Method 2
measuring the time the light is reflected, and arriving at the object... a method of measuring ToF using a sensor capable of measuring a light arrival time in units of pico seconds, e.g., a single photon avalanche diode (SPAD)
Data Source
AI summary
A depth map generation apparatus includes: a sensor array including a plurality of sensors configured to detect a received light arrival time; a control circuit array; and a processor configured to divide an acquisition time of at least one of the plurality of sensors into a first time window and a second time window; based on a detection result of the at least one sensor in the first time window and the second time window, generate first sub-frame information; based on the first sub-frame information, determine in which one of the first time window and the second time window an object is detected; generate second sub-frame information from the one of the first time window and the second time window in which the object is detected; and provide the first sub-frame information and the second sub-frame information for generation of an image frame.