Human Body Tracking Using KCF and Multi-Device Features
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing human body tracking methods in images suffer from poor accuracy due to changes in the position of human bodies in real-time, leading to inconsistencies across different frames.
Innovation Solution
The method employs a kernelized correlation filters (KCF) tracking algorithm to calculate second position information of human bodies in current frames, combined with pedestrian features from historical images captured by multiple devices, to determine the most accurate target position information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional detection methods are used to obtain human body position in each frame, then the tracking process is simple, but the tracking accuracy deteriorates due to real-time position changes
Solution Approach 1:
The system pre-calculates and stores position information of human bodies in previous frames before current frame processing. This preliminary preparation enables the KCF algorithm to efficiently compare and match positions across frames, improving tracking accuracy without adding complex real-time processing requirements.
Solution Approach 2:
The patent introduces KCF (Kernelized Correlation Filters) as an intermediary algorithm that bridges traditional detection and accurate tracking. The KCF algorithm acts as a mediator that utilizes position information from multiple frames to determine accurate current position, resolving the contradiction between simple processing and high accuracy.
2Measurement precision
If KCF tracking algorithm is used to calculate position information, then tracking accuracy improves, but computational complexity increases
Solution Approach 1:
The system applies KCF tracking selectively to human bodies that are detected in previous frames, rather than processing all objects in every frame. This partial application of the computationally intensive KCF algorithm reduces overall power consumption while maintaining high accuracy for tracked targets.
Solution Approach 2:
Position information from previous frames is pre-calculated and stored before KCF processing. This preliminary preparation reduces the real-time computational burden during KCF execution, lowering power consumption while maintaining high position calculation accuracy.
3Reliability
If multiple photographing devices are used to capture historical images, then tracking reliability improves, but system complexity increases
Solution Approach 1:
The patent merges data from multiple photographing devices by collecting position information from all devices and processing it through a unified KCF tracking framework. This merging approach improves tracking reliability through multiple data sources while managing system complexity through centralized processing logic.
Solution Approach 2:
The system design makes the tracking apparatus multi-functional by enabling it to process images from multiple photographing devices through the same KCF algorithm framework. This universal approach allows reliable multi-device tracking without requiring separate processing systems for each device.
Data Source
Figure 1
Figure 2~3
Figure 4~5
AI summary
Embodiments of the present invention provide a human body tracing method, apparatus, and device, and a storage medium. The method includes: obtaining a current frame image captured by a target photographing device at a current moment; detecting each human body in the current frame image to obtain first position information of the each human body in the current frame image; calculating , by using a preset tracing algorithm, second position information of a first human body in the current frame image, where the first human body is traced in a previous frame image ahead of the current frame image; determining target position information of the each human body in the current frame image according to the second position information of the first human body in the current frame image, the first position information of the each human body in the current frame image, and pedestrian features of all traced pedestrians stored in a preset list; where the pedestrian features of all the traced pedestrians stored in the preset list are determined according to historical images captured by a plurality of photographing devices. The embodiments of the present invention improve the accuracy of human body tracing in an image.