DOAJ Open Access 2021

Stereo Matching Method Based on Multi-Scale and Multi-Level Features Fusion

WANG Jinhe, CHE Zhilong, ZHANG Nan, MENG Fanyun, SU Cuili, TAN Hao

Lihat Sumber DOI

Abstrak

The stereo matching method based on Convolutional Neural Network(CNN) does not make full use of the feature map information of each level in the image,resulting in the poor feature extraction performance in the ill posed region in the image.This paper proposes a stereo matching method based on multi-scale and multi-level features.A pooled pyramid layer is designed at the front end of the CNN model with double tower structure to extract the multi-scale low-level structural features of the image.In the back end of the network model,the high-level semantic features of the last three layers of network are fused to extract image features,and the disparity map is output after similarity measurement of image features.The experimental results on KITTI 2015 dataset show that compared with the LUO and Anita methods,the proposed method reduces the pixel error accuracy from 14.65% and 8.30% to 8.02%,and can obtain a disparity map with better detail information.

Topik & Kata Kunci

Computer engineering. Computer hardware Computer software

Penulis (1)

WANG Jinhe, CHE Zhilong, ZHANG Nan, MENG Fanyun, SU Cuili, TAN Hao

Format Sitasi

APA MLA BibTeX

Hao, W.J.C.Z.Z.N.M.F.S.C.T. (2021). Stereo Matching Method Based on Multi-Scale and Multi-Level Features Fusion. https://doi.org/10.19678/j.issn.1000-3428.0056715

Akses Cepat

Lihat di Sumber doi.org/10.19678/j.issn.1000-3428.0056715

Informasi Jurnal

Tahun Terbit: 2021
Sumber Database: DOAJ
DOI: 10.19678/j.issn.1000-3428.0056715
Akses: Open Access ✓