Pedestrian detection method and system based on vehicle-mounted infrared video

A pedestrian detection and infrared technology, applied in the field of computer vision for pedestrian detection, can solve problems such as the influence of pedestrian accuracy

Active Publication Date: 2020-09-29
SOUTH CENTRAL UNIVERSITY FOR NATIONALITIES
View PDF4 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, if the pedestrians in the video are occluded, the accuracy of pedestrian detection is still affected

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 more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Pedestrian detection method and system based on vehicle-mounted infrared video
  • Pedestrian detection method and system based on vehicle-mounted infrared video
  • Pedestrian detection method and system based on vehicle-mounted infrared video

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0064] see figure 1 As shown, embodiment 1 of the present invention provides a kind of pedestrian detection method based on vehicle-mounted infrared video, comprises the following steps:

[0065] S1. Obtain the infrared video through the vehicle-mounted infrared detector, and extract the color feature and gradient feature of the infrared video by using the integral map and channel features;

[0066] S2. Using the adaptive enhanced iterative algorithm Adaboost to train and screen the features extracted in step S1, detect pedestrians in the infrared video images, and obtain preliminary detection results;

[0067] S3. Using the Hungarian algorithm to calculate the preliminary detection results obtained in step S2, and obtain the initial trajectory sequence formed by the association of pedestrians in all images;

[0068] S4. Using the initial trajectory sequence obtained in step S3 as an initial value, the Mancles algorithm is used to calculate optimal trajectory correlation info...

Embodiment 2

[0070] On the basis of Embodiment 1, step S1 specifically includes the following steps:

[0071]The infrared video is obtained by the vehicle-mounted infrared detector, and the image of the infrared video is zoomed into an image pyramid; on different scales, using the integral map, the three color channel features L(I) of each frame image I after zooming, U(I), V(I), 6 gradient direction channel features G(I) 1[Θ=i] with angle Θ=i, 1 gradient amplitude channel feature||G(I)|| Extract respectively, and then perform normalization processing to obtain the integrated feature vector F representing the color feature and gradient feature of the frame image. The calculation formula of the integrated feature vector F is:

[0072] F={L(I),U(I),V(I),||G(I)||,G(I)·1[Θ=i]},

[0073]

Embodiment 3

[0075] On the basis of embodiment 2, step S2 specifically includes the following steps:

[0076] Let the input n training samples be: {(x 1 ,y 1 ),(x 2 ,y 2 ),…,(x n ,y n )}, where x i is the input training sample, y i ∈{0,1} represent positive samples and negative samples respectively, where the number of positive samples is l, the number of negative samples is m, n=l+m, l, m, and n are all positive integers; the adaptive enhanced iterative algorithm is adopted Adaboost, initialize the weight of each sample, train a weak classifier for detecting whether there are pedestrians in the image for each channel feature, calculate the weighted error rate of the weak classifier, adjust the weight according to the weak classifier with the smallest error rate, and iterate The selection process of the weak classifier is to synthesize the selected weak classifier into a strong classifier to detect whether there are pedestrians in the image; the strong classifier is used to train an...

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

No PUM Login to view more

Abstract

The invention discloses a pedestrian detection method and system based on vehicle-mounted infrared video, and relates to the field of computer vision for pedestrian detection. The method includes the following steps: acquiring infrared video through a vehicle-mounted infrared detector, extracting color features and gradient features of the infrared video by using the integral map and channel features; using the adaptive enhanced iterative algorithm Adaboost, training and screening the extracted features, and detecting Pedestrians in the infrared video image, get the preliminary detection results; use the Hungarian algorithm to calculate the preliminary detection results, and get the initial trajectory sequence formed by the association of pedestrians in all images; take the initial trajectory sequence as the initial value, and use Mancles Algorithm to calculate the optimal trajectory associated information to obtain the final tracking trajectory. The invention combines detection and tracking, can detect pedestrians who are covered during driving, and improves the accuracy of pedestrian detection.

Description

technical field [0001] The invention relates to the field of computer vision for pedestrian detection, in particular to a pedestrian detection method and system based on vehicle-mounted infrared video. Background technique [0002] According to the classification of detection methods, pedestrian detection can be divided into methods based on motion characteristics, methods based on template matching and methods based on statistical learning. Compared with the previous two pedestrian detection methods, the method based on statistical learning has the advantages of high detection accuracy and better robustness, and is the focus and hot spot of pedestrian detection research today. The method based on statistical learning uses a large number of pedestrian samples to extract information such as grayscale, edge, texture, and color of the target, and uses learning methods to construct a pedestrian detection classifier. Learning methods mainly include adaptive enhanced iterative al...

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
no application Login to view more
Patent Type & Authority Patents(China)
IPC IPC(8): G06K9/00G06K9/46G06K9/62G06T7/246
CPCG06T7/246G06T2207/10048G06T2207/10016G06T2207/30241G06T2207/30196G06V40/103G06V10/56G06F18/2148G06F18/22G06F18/24
Inventor 刘李漫刘佳谌先敢刘海华
Owner SOUTH CENTRAL UNIVERSITY FOR NATIONALITIES
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap