A Pedestrian Detection Method Based on Deep Learning in Complex Background

Abstract

The invention discloses a complex background pedestrian detection method based on deep learning. The main steps are as follows: build an SSD target detection framework and modify the SSD network model: add an Inception component in the middle of a specific convolution layer, and use the Concatenation operation to convert the specific convolution layer Perform fusion to obtain a new feature extraction layer, and build a new feature extraction network based on the new feature extraction layer; add an Inception component to the back end of the new feature extraction layer, and use Pooling and Concatenation operations to combine the extracted features layer by layer with subsequent The feature extraction layer is fused; the network obtained in the above steps is used as the training network, and the pedestrian data set is used for training; the relevant threshold is set, the corresponding test set is used for testing, and the detection result is output. The invention fully excavates the background information in the image by deepening and widening the neural network, which not only improves the recall rate of the SSD framework in pedestrian detection tasks, but also enhances the robustness of the framework, and improves the performance of the framework in complex backgrounds. pedestrian detection performance.

Description

technical field [0001] The invention relates to the technical fields of computer vision and artificial intelligence, in particular to a method for detecting pedestrians in complex backgrounds based on deep learning. Background technique [0002] Pedestrian detection is to judge whether there is a pedestrian in the input image or video sequence, and determine its position. Pedestrian detection is a branch of object detection, and it is also a research hotspot and difficulty in the field of object detection. It has a wide range of applications in the fields of artificial intelligence such as automatic driving, video surveillance, and intelligent robots. At the same time, pedestrian detection is also the premise and basis of many computer vision tasks, such as pedestrian structuring, pedestrian behavior analysis, and pedestrian re-identification. It is necessary to detect pedestrians in the input data before subsequent work can be performed. Therefore, effective pedestrian det...

AI Technical Summary

Problems solved by technology

[0006] The purpose of the present invention is to overcome the shortcomings and deficiencies of the prior art, and provide a complex background pedestrian detection method based on deep learning. Disadvantages, enhance the robustness of the SSD framework, and improve the pedestrian detection performance of the framework in complex backgrounds

Images