Pedestrian recognition method based on light field camera and deep learning

Abstract

The invention discloses a pedestrian recognition method based on a light field camera and a deep learning technology. The pedestrian recognition method comprises the following steps: (1) acquiring a plurality of pedestrian images by utilizing the light field camera; (2) processing the original pedestrian image obtained in the step (1) by utilizing Lytro desktop software to obtain a color pedestrian image and a depth pedestrian image; (3) preprocessing the color image and the depth image obtained in the step (2), classifying the images into a unified size, and dividing the images into positiveand negative samples to obtain a light field image data set; (4), initializing the model based on deep learning; (5) constructing a joint convolution feature of the color image and the depth image; (6) utilizing the method in the step (4), utilizing an existing ResNet50 image classification model trained on an Image Net data set, and utilizing model parameters to extract convolution features of atraining sample; and (7) repeatedly training the neural network according to the features obtained in the step (6) to obtain a new classification model. The accuracy and robustness of the pedestrian detection method can be effectively improved.

Description

technical field [0001] The invention belongs to the field of computer vision, in particular to a pedestrian recognition method based on a light field camera and deep learning. Background technique [0002] Pedestrian recognition is an important part of the computer vision research field, and it has played a very important role in the fields of intelligent transportation, video surveillance, artificial intelligence and automatic driving. In recent years, based on the rapid development of computer hardware equipment and new camera technology, the industry has put forward more stringent requirements for the performance and accuracy of pedestrian recognition. [0003] Due to the vigorous development of autonomous driving technology in recent years, the accuracy of pedestrian recognition is particularly important. Because pedestrians have the characteristics of rigid objects and non-rigid objects, such as the variability of pedestrian shooting angles, the existence of factors su...

Images