Pedestrian real-time detection method in complex environment based on deep learning

Abstract

The invention provides a pedestrian real-time detection method in a complex environment based on deep learning. The pedestrian real-time detection method comprises the following steps: S1, establishing a detection model based on a YOLO algorithm; S2, selecting a plurality of pedestrian images in a complex environment in a color hot picture library, establishing a training data set and a test dataset, inputting the training data set into the detection model, and training the detection model; S3, inputting the test data set into the trained detection model, outputting detection results of pedestrian targets in an infrared thermal image and an RGB color image of the test data set, and screening the detection results through a non-maximum suppression method; and S4, comparing with a YOLOv3, YOLO-tiny detection algorithm, and verifying the detection precision and the detection speed.

Description

technical field [0001] The invention belongs to the field of target recognition, in particular to a method for real-time detection of pedestrians in complex environments based on deep learning. Background technique [0002] Pedestrian detection has attracted extensive attention from the computer vision community due to its wide applications in areas such as driver assistance (autonomous driving vehicles), robotics, person re-identification, video surveillance, and pedestrian behavior analysis. At present, compared with traditional methods, deep learning technology has achieved good results in the field of pedestrian detection. However, in some complex natural environments, the results of detection tasks relying only on visible spectrum images or infrared spectrum images are not accurate enough. Pedestrian targets will encounter many challenges in complex environments, such as: smog, rain, dust, dim light, etc. RGB color image is rich in spectral information, and can reflec...

AI Technical Summary

Problems solved by technology

However, in some complex natural environments, the results of detection tasks that only rely on visible spectrum images or infrared spectrum images are not accurate enough

Pedestrian targets will encounter many challenges in complex environments, such as: smoke, rain, dust, dim light, etc.

RGB color image is rich in spectral information, and can reflect the details of the scene under certain lighting conditions, but it is difficult to detect the target when the visibility is poor; infrared thermal image is a kind of thermal radiation image, and the gray level is determined by the difference between the observed target and the background. The determination of the temperature difference between them usually lacks structural information, therefore, the target is easily mixed with the background, resulting in false detection and missed detection

Images