Infrared image pedestrian target detection method based on improved YOLOv5

Abstract

The invention provides an infrared image pedestrian target detection method based on improved YOLOv5, and the method comprises the following steps: expanding and iterating a shallow CSP module of a feature extraction network, adding an improved attention module into a residual block, adding a multi-scale target detection layer, downloading and processing a KAIST data set, and constructing a training set, a verification set and a test set for model training; finally, sending the preprocessed KAIST data set into the constructed infrared pedestrian target detection model based on the improved YOLOv5 to carry out model training, testing and evaluation. In the model construction stage, the expanded CSP and the introduced attention mechanism are more beneficial to extraction of pedestrian features, and the added detection layer is beneficial to detection of long-distance small targets. In the training stage, infrared images which do not contain pedestrian targets in a sent data set are deleted, so that model training is prevented from being interfered, and network convergence is accelerated. In the evaluation stage, the accuracy and the speed of the model are optimized by adjusting the width and the depth of the model so as to meet the requirements of practical application.

Description

technical field [0001] The invention relates to the technical field of infrared image pedestrian target detection, in particular to an infrared image pedestrian target detection method based on improved YOLOv5. Background technique [0002] Pedestrian detection requires accurate judgment of whether the input image or video contains pedestrians, and gives the spatial coordinate information of pedestrians in the image, which is widely used in intelligent monitoring, area investigation, human behavior understanding, automatic driving and other fields. Compared with traditional pedestrian detection tasks based on visible light, infrared imaging has strong anti-interference ability, is less affected by light and bad weather, and has the ability to work around the clock. However, the contrast of the infrared image is low, the texture features are weak, and the interference is large. Under the influence of strong noise and similar background, the pedestrian target becomes a weak ta...

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to provide an infrared image pedestrian target detection method based on improved YOLOv5 to solve the problems of insufficient accuracy in the existing infrared image detection method in the background technology

Images