Library floating personnel target detection and multi-target tracking method and management system
A multi-target tracking and target detection algorithm technology, applied in the field of computer vision, can solve the problems of high cost of sensor networks and inability to accurately detect densely populated areas
Pending Publication Date: 2022-01-11
TONGJI UNIV
View PDF0 Cites 1 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
[0010] (1) In view of the high cost of the sensor network and the inability to accurately detect densely populated areas, the data set is first screened to cater to the actual environment of the library. After setting the model training parameters and training environment, the model is trained and tested. Analyzing training results
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
specific Embodiment approach
[0076] The concrete implementation process of the present invention comprises target detection algorithm, multi-target tracking algorithm and library intelligent management system construction three aspects specific implementation modes are as follows:
[0077] 1. Target detection algorithm
[0078] (1) Training data
[0079] The COCO (Common Object in COntext) dataset is a dataset provided by Microsoft for image recognition. The COCO dataset has at least 80 categories, including people, bicycles, notebooks, mice, books, and more. The COCO dataset is huge. Taking COCO2017 as an example, its training set contains at least 110,000 pictures, and the test set contains 5,000 pictures, and its total size after compression is about 18GB.
[0080] Due to the relevant hardware limitations of the computer used for model training, the training data set used by the real-time video detection model of the present invention is a subset of the COCO data set, including 5000 training picture...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More PUM
Login to View More Abstract
The invention provides a library floating personnel target detection and multi-target tracking method and management system. The method comprises steps of carrying out target detection on the video by adopting a target detection method based on a Yolov5 model, training the model by screening a data set and adjusting proper model training parameters, and carrying out model testing, target detection and related data analysis; improving a Deep SORT algorithm, and carrying out multi-target tracking on a target detection result; and constructing a library intelligent management system based on physical space data analysis according to the target detection and multi-target tracking results. The problems that a sensor network is high in cost, incapable of accurately detecting a crowded area and high in dependence on continuity of a target are solved, real-time query service of library use conditions which cannot be provided from the Internet or telephone service is provided for students, personalized service and intelligent management of the library are implemented for library managers according to real-time personnel flow conditions.
Description
technical field [0001] The invention relates to the field of computer vision, in particular to a target detection and multi-target tracking method for floating personnel in a library. Background technique [0002] Regarding the object detection in the library environment, it is necessary to count the people or objects in the library learning area, so it is necessary to perform object detection in the learning area. Currently there is a technology that uses infrared intrusion detectors to count the number of people in a room. Although this people counting method has the advantages of low cost, fast response, and small amount of calculation, it is mainly aimed at counting people at the entrance and exit of one-way traffic scenes, and the detection accuracy of pedestrians walking side by side is low. [0003] Using radar is also a good way to detect indoor personnel. Its detection accuracy can reach more than 95%, and the radiation is small, the response speed is fast, and the...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More Application Information
Patent Timeline
Login to View More Patent Type & Authority Applications(China)
IPC IPC(8): G06V10/774G06V20/52G06V20/40G06N3/08
CPCG06N3/084G06F18/214
Inventor 程久军许国望陈嘉捷熊永红
Owner TONGJI UNIV