Unlock instant, AI-driven research and patent intelligence for your innovation.

Night road pedestrian detection method based on YOLOv3-tin-DB and transfer learning

A pedestrian detection and migration learning technology, applied in neural learning methods, image enhancement, instruments, etc., can solve the problems of large contrast difference and little color information, and achieve the effect of improving the recognition rate and increasing the display effect.

Pending Publication Date: 2022-05-31
KUNMING UNIV OF SCI & TECH
View PDF0 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0006] The technical problem to be solved by the present invention is to provide a nighttime road pedestrian detection method based on YOLO v3-tiny-DB and transfer learning, which is used to solve the problems of large contrast difference and less color information of the image formed by the vehicle-mounted camera in the nighttime environment, so as to enhance The display effect of the image formed by the on-board camera improves the recognition rate of pedestrian detection
Finally, load the training weights and transplant them to the local assisted driving platform to improve the pedestrian recognition rate of night road images

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
  • Night road pedestrian detection method based on YOLOv3-tin-DB and transfer learning
  • Night road pedestrian detection method based on YOLOv3-tin-DB and transfer learning
  • Night road pedestrian detection method based on YOLOv3-tin-DB and transfer learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0031] Embodiment 1: as Figure 1-9 As shown, a nighttime road pedestrian detection method based on YOLO v3-tiny-DB and migration learning, the specific steps are:

[0032] Step1: Firstly, high-definition vehicle cameras were used to collect nighttime road pedestrian images in various streets, and 19480 nighttime images were obtained.

[0033] Step2: Preprocess the nighttime images. Firstly, use the improved limited contrast histogram equalization algorithm to obtain brightness images from 19,480 nighttime images. Then, the original nighttime images and the processed brightness images are processed by Gaussian pyramid and Laplacian pyramid. The final image is obtained by fusion, and the original night image and the final image are cross-stacked to establish a nighttime road pedestrian dataset.

[0034] Step3: Import the nighttime road pedestrian detection data set into the target detection network of YOLO v3-tiny-DB, adjust the network structure and the input size of the nigh...

Embodiment 2

[0041] Embodiment 2: the core of the present invention is to provide a nighttime road pedestrian detection method based on YOLO v3-tiny-DB and transfer learning, the first can improve the contrast and color scale of the nighttime road image of the visual sensor, so that the processed image looks It is clearer and helps the model capture the characteristics of pedestrians. The second is that based on dense connections, it can improve the detection accuracy of YOLO v3-tiny detection network for pedestrians, and improve the safety of pedestrian detection at night for assisted driving. Third, transplanting the training weights to the local assisted driving platform through transfer learning can improve the effect of pedestrian detection at night.

[0042] The nighttime road image pedestrian detection method used in the present invention is the YOLO v3-tiny-DB model, which is improved based on the YOLO target detection network to realize target detection.

[0043] The main task of ...

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 relates to a night road pedestrian detection method based on YOLO v3-tin-DB and transfer learning, and belongs to the technical field of computer image processing. A method in the field of computer image processing is utilized, an improved contrast-limited histogram equalization algorithm is used for a night image to obtain a brightness image, and then the original night image and the processed brightness image are fused through a Gaussian pyramid and a Laplacian pyramid to obtain a final image. And the original night image and the final image are crossly stacked to establish a night road pedestrian data set. A brand new YOLO v3-tiny-DB pedestrian detection network is designed, and a night road pedestrian data set is trained through the YOLO v3-tiny-DB network to obtain a training weight. And finally, the training weight is loaded and transplanted to a local auxiliary driving platform to improve the pedestrian recognition rate of the night road image. Compared with the prior art, the display effect of the image formed by the vehicle-mounted camera is improved, and the recognition rate of the detection model for pedestrians on the road at night is improved.

Description

technical field [0001] The invention relates to a nighttime road pedestrian detection method based on YOLO v3-tiny-DB and transfer learning, belonging to the technical field of computer image processing. Background technique [0002] Benefiting from the rapid development of deep learning methods in recent years, many computer vision applications have been developed to design advanced driver assistance systems (ADAS) and connected autonomous vehicles (CAV). These applications mainly focus on object detection, object classification, object recognition, semantic segmentation, motion estimation and surveillance systems. However, most of the available computer vision applications are based on visible light cameras and thus can only be used in normal light and clear weather conditions, making most recent models unsuitable for nighttime images. Traffic safety statistics show that 51.1 percent of fatal crashes in the United States occur at night (from 6 p.m. to 6 a.m.), especially ...

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
IPC IPC(8): G06T7/00G06T5/40G06T5/00G06N3/04G06N3/08
CPCG06T7/0002G06T5/40G06N3/08G06N3/045G06T5/90Y02T10/40
Inventor 曾凯沙梦洲沈韬刘英莉陈敏
Owner KUNMING UNIV OF SCI & TECH