Logarithmic Histogram Bins for TOF Sensor Silicon Area Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Time-of-Flight (TOF) imaging systems, such as LIDAR, face challenges in efficiently storing and processing large dynamic ranges of returned photon flux due to the high data rates and silicon area requirements of time-domain histograms, especially when dealing with massive distance ranges.
Innovation Solution
The method involves using a histogram with bins sized based on a logarithmic function of the respective distance range, along with digital-to-time converters and counters in image processing circuitry, to generate and store counts associated with distances, reducing the silicon area needed for storage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a time-domain histogram with many bins is used to store multiple time slices or depth points, then the imaging system can capture a large dynamic range of returned photon flux, but the silicon area cost becomes considerable
Solution Approach 1:
The histogram is segmented into multiple depth points, where each depth point has its own histogram with a limited number of bins. This segmentation allows the system to handle large dynamic ranges by dividing the photon flux data across multiple depth points, each with manageable bin counts, thereby reducing the total silicon area required compared to a single large histogram
Solution Approach 2:
The patent introduces a new dimension (depth point index) to organize histogram data. Instead of using a single flat histogram structure, the data is organized as a two-dimensional structure with depth points as one dimension and histogram bins as another. This dimensional transformation enables efficient memory addressing and reduces the silicon area by allowing shared storage structures across multiple depth points
2Productivity
If Time-to-Digital Converters (TDCs) with direct histogram outputs are used, then the system can capture time-domain histograms at high data rates, but the data rate becomes very high requiring considerable storage resources
Solution Approach 1:
The high data rate histogram output from TDCs is segmented into multiple depth points, each with a reduced number of bins. This segmentation distributes the storage burden across multiple smaller structures, maintaining the high data rate capability while reducing the total storage area required compared to a single large histogram
Solution Approach 2:
The patent performs preliminary organization of histogram data into depth-point-specific structures at the time of data generation. By pre-organizing the data into manageable depth point segments with appropriate bin allocations, the system prepares the data for efficient storage and processing, reducing the area required for subsequent storage operations
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 significantly reduces the silicon area required for storing histogram data, achieving area savings of over 65% and enabling efficient processing of large dynamic ranges in TOF imaging systems.
Implementation Method 1
generating, using the image sensor, an image data stream based on reflections from the one or more objects
Implementation Method 2
illumination circuitry, which, in operation, illuminates one or more objects in an environment around the device
Data Source
AI summary
A time-of-flight (TOF) imaging system includes illumination circuitry, such as a laser, one or more sensors, such as SPAD arrays, and image processing circuitry. The illumination circuitry illuminates one or more objects in an environment around the TOF imaging system. The sensors generate an image data stream based on reflections from the one or more illuminated objects, and possibly based on reflections from one or more reflectors. The image processing circuitry generates counts associated with distances based on the image data stream and possibly a reflection data stream and stores the generated counts in a histogram using a plurality of bins. Each of the plurality of bins stores counts associated with a respective distance range. A size of a bin in the plurality of bins is a function of the respective distance range, and may be based on a logarithmic function of the distance associated with the bin.