Two-Stage Video Frame Throttling for ALPR Stability
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Automatic License Plate Recognition (ALPR) systems face challenges in efficiently processing video frames from multiple lanes of vehicle traffic, leading to potential system instability and resource overload, which can result in missed license plate detections due to excessive frame processing.
Innovation Solution
A two-stage dynamic throttling method is implemented in the ALPR system, where the first stage drops a subset of video frames based on general policies and the second stage further drops frames based on lane-specific conditions to control frame input into the processor, ensuring stable processing and preventing resource overload.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If all video frames are processed by the processor, then complete license plate detection is achieved, but system resource overload and instability occur
Solution Approach 1:
The frame dropping process is segmented into two distinct stages: first stage dropping based on global frame rate policies, and second stage dropping based on lane-specific conditions. This segmentation allows the system to manage processing load more effectively by dividing the decision-making process into hierarchical levels, preventing resource overload while maintaining detection capability.
Solution Approach 2:
Instead of processing all frames or none, the system applies partial action by selectively dropping subsets of frames at two different stages. The first stage drops frames globally, and the second stage drops additional frames based on lane conditions, achieving a balanced approach that maintains system stability while preserving sufficient frames for reliable license plate detection.
2Reliability
If frame processing is reduced to prevent resource overload, then system stability is improved, but license plate detection accuracy may deteriorate
Solution Approach 1:
The first stage of frame dropping is performed as a preliminary action before the second stage. By pre-filtering frames based on global frame rate policies, the system reduces the overall processing load early in the pipeline, preventing resource overload before frames reach the processing stage, while still allowing subsequent lane-specific filtering to preserve critical detection frames.
Solution Approach 2:
The two-stage dropping system acts as an intermediary mechanism between raw video input and processor analysis. This intermediary structure allows the system to buffer and regulate frame flow, preventing direct overload of the processor while maintaining a sufficient supply of frames for accurate license plate detection through coordinated dropping policies at both stages.
3Device complexity
If a single-stage frame dropping method is used, then device complexity is reduced, but control precision over frame input is insufficient
Solution Approach 1:
The system transitions from a single-dimensional frame dropping approach to a two-dimensional hierarchical structure. The first stage operates globally across all lanes, while the second stage operates specifically per lane, adding a dimensional layer of control. This dimensional expansion enables precise frame selection without proportionally increasing overall system complexity, as each stage builds upon the previous one.
Data Source
AI summary
An n-lane automatic license plate recognition system (where n is an integer greater than one) and a method for dropping video frames in this system is disclosed. The method includes receiving, at an input of a first stage of a two-stage restrictor, a plurality of video frames. In the first stage of the two-stage restrictor, a subset of the plurality of video frames is dropped. An output of the first stage is coupled to an input of a second stage of the two-stage restrictor such that receiving and non-receiving of individual ones the plurality of video frames by the second stage is controlled by the first stage. In the second stage after the first stage, a further subset of the plurality of video frames is dropped to effect controlled frame input into a processor that performs at least one recognition analytic on remaining video frames of the plurality of video frames. The dropping of the further subset is based on at least one drop condition including a lane-specific drop condition.