Motion Detection System Using Luminance Extraction and Resolution Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing motion-detection systems face challenges in efficiently detecting moving objects while minimizing power consumption and addressing high computation and storage requirements.
Innovation Solution
A motion-detection system comprising a camera and a storage device that operates in low-power mode by generating a luminance image, performing resolution reduction, and switching to normal mode only when motion is detected, using a neural-network processing unit to identify and mask non-predefined objects, thereby reducing computation and power consumption.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the system operates in normal mode to ensure accurate motion detection, then detection reliability is improved, but power consumption increases
Solution Approach 1:
The system dynamically switches between low-power mode and normal mode based on motion detection needs. The processor operates in low-power mode during static periods and transitions to normal mode only when motion is detected, making the operational state adaptive rather than fixed. This resolves the contradiction by allowing the system to maintain reliability when needed while minimizing power consumption during idle periods.
Solution Approach 2:
The system employs periodic motion detection cycles where the processor alternates between sleeping (low-power mode) and active (normal mode) states. Motion detection is performed periodically at intervals, with the processor waking up to check for motion and returning to sleep mode if no motion is detected. This periodic operation pattern reduces overall power consumption while maintaining adequate detection reliability.
2Measurement precision
If the system performs full image processing to ensure accurate object identification, then detection precision is improved, but computation time increases
Solution Approach 1:
The image processing is segmented into two stages: a first stage using a lightweight neural network model for quick preliminary classification, and a second stage using a more complex model only for objects that pass the first filter. This segmentation allows the system to achieve high identification precision for relevant objects while minimizing computation time by avoiding full processing of all objects in the scene.
Solution Approach 2:
The system applies partial processing by using a simplified neural network model for initial object classification and reservation judgment. Only objects that meet certain criteria (e.g., high confidence of being a target object) trigger the more computationally intensive second-stage processing. This partial action approach achieves sufficient precision for the application while significantly reducing overall computation time.
3Loss of information
If the system stores and processes high-resolution color images to maintain image quality, then image quality is improved, but memory usage increases
Solution Approach 1:
The system extracts only the necessary information from full-color images for motion detection and object identification purposes. Specifically, it extracts luminance information and key object features rather than processing and storing complete high-resolution color data. This extraction approach maintains sufficient image quality for detection tasks while dramatically reducing memory usage requirements.
Solution Approach 2:
The system uses disposable, low-cost neural network models that process images in a single pass without requiring storage of intermediate high-resolution data. The lightweight models operate on compressed or downsampled image data, performing classification and detection tasks without needing to retain the original high-quality images in memory, thus reducing memory requirements while maintaining detection accuracy.
Data Source
AI summary
A motion-detection system including a camera and a storage device is provided. The camera captures a color image and transmits the color image to a pre-processor. In a low-power mode, the pre-processor retrieves the luminance value of each pixel in the color image to generate a luminance image and performs resolution reduction on the luminance image to generate an input image. The motion-identifying device identifies a motion-object block within the input image and calculates the amount of movement of the motion-object block. When the amount of movement reaches an identification threshold, the motion-detection system enters a normal mode from the low-power mode. In the normal mode, when an object-identifying model determines that an object in the motion-object block is not a predefined object, the motion-detection system enters the low-power mode.