Method for detecting pedestrians on road

Abstract

The invention relates to the field of computer vision, in particular to a method for detecting pedestrians on a road. The problems that pedestrian detection is greatly influenced by scenes, the detection rate is low, and the false alarm rate is high are solved. The method comprises the following steps: step 1, collecting videos of pedestrians on a road, and extracting video frame images; step 2, preprocessing the video frame image; step 3, HOG features and Haar features of the image are extracted; 4, dimensionality reduction and feature fusion are carried out on the extracted feature vectors; 5, designing a proper classifier; and step 6, acquiring an image to be detected, performing steps 2-4 on the image, inputting the image into the classifier for detection, and outputting a detection result. Preprocessing such as graying and histogram equalization is carried out on a video frame image, so that subsequent feature extraction and classification of the image are facilitated, and meanwhile, the calculation amount is also reduced; by extracting a plurality of features, the description of pedestrian features is clearer, the robustness of pedestrian detection is improved, the accuracy of pedestrian detection is increased, and the false detection rate is reduced.

Description

Technical field [0001] The present invention relates to the field of computer vision, in particular to a pedestrian detection method on the road. Background [0002] Pedestrian detection, which includes aspects such as pedestrian self-marking classification and pedestrian tracking, involves many core topics in the fields of computer vision, pattern recognition and artificial intelligence, and is a challenging and difficult problem. At present, in many literatures, a variety of methods for pedestrian detection have been proposed, and the commonly used methods are methods based on motion characteristics, methods based on shape information, methods based on pedestrian models, stereo vision methods, neural network methods, wavelet transforms and methods for support vector machines, etc., which have their own advantages and disadvantages. For example: methods based on motion characteristics can only detect pedestrians in motion, not prohibited people; The method based on shape informa...

AI Technical Summary

Problems solved by technology

[0004] The purpose of the present invention is to provide a pedestrian detection method on the road, which is used to solve the problems that pedestrian detection is greatly affected by the scene, the detection rate is low, and the false alarm rate is high, so as to improve the detection rate and reduce the false detection rate

Images