Image Sensor ROI Motion Detection for Low-Bandwidth Vision
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image sensors face challenges in capturing high-speed, high-resolution images with large bandwidth and power consumption, particularly in applications like surveillance and robotic vision, due to the need for external processor computing and raw image transfer.
Innovation Solution
An image sensor integrated with a processing-in-sensor (PIS) circuit that generates frame-difference signals and identifies regions of interest (ROI) without external processing, reducing power consumption and latency by performing motion detection directly on the sensor.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If high frame-rate images with high spatial resolution are captured, then image quality is improved, but bandwidth and power consumption increase
Solution Approach 1:
The patent divides the image processing task by identifying and separating regions of interest (ROI) from non-interest areas. Instead of processing the entire high-resolution image, the system segments only the relevant portions that contain motion information, thereby reducing the data volume for processing and transmission while maintaining image quality where needed.
Solution Approach 2:
The patent extracts only the essential information from the image data by detecting motion and identifying ROIs. The system takes out and processes only the critical motion detection tasks within the sensor array, separating this function from the full image processing pipeline, which reduces computational load and power consumption while preserving important visual information.
2Productivity
If external processor computing and raw image transfer are used, then processing capability is improved, but latency and bandwidth requirements increase
Solution Approach 1:
The patent merges the image sensing and motion detection functions into a single integrated system. The processing circuit is combined with the pixel array to perform motion detection locally at the sensor level, eliminating the need for separate external processors and reducing the data transfer distance, thereby decreasing latency while maintaining processing capability.
Solution Approach 2:
The patent performs preliminary motion detection and ROI identification directly at the sensor level before data transmission to external processors. This preliminary action filters and prepares the data in advance, reducing the amount of data that needs to be transmitted and processed externally, thereby reducing latency and bandwidth requirements.
3Use of energy by moving object
If processing-in-sensor is implemented, then power consumption is reduced, but device complexity increases
Solution Approach 1:
The patent implements a multi-functional pixel structure where each pixel serves both as a light sensor and as a motion detection unit. The processing circuit performs multiple functions including frame difference calculation, motion detection, and ROI identification, thereby reducing the need for separate dedicated components and managing device complexity through functional integration.
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 high-speed, low-power, and real-time ROI motion detection, reducing the need for external processing and minimizing bandwidth and memory requirements.
Implementation Method 1
a pixel array and an analog-to-digital converter generate raw images of frames and further generate frame-difference signals by comparing light intensity changes
Data Source
AI summary
A method is provided to operate an image sensor, including the following steps: comparing a first threshold with multiple first frame-difference signals to generate a first region of interest (ROI) address, wherein each of multiple first frame-difference signals is generated by a corresponding pixel, in multiple pixels, operating during a first time frame and a second time frame, wherein multiple pixels are arranged in multiple columns and multiple rows; identifying edge columns and edge rows in the first ROI address; adjusting, according to the edge columns and the edge rows, the first ROI address to generate a second ROI address; and outputting a ROI image according to the second ROI address and multiple second frame-difference signals generated by multiple pixels operating during the second time frame and a third time frame after the second time frame.