Open AccessImproved MSER Pedestrian Detection Algorithm based on TOF Camera
Author(s) -
ZiJin Wang,
Yiqin Zhao,
Chao Zhao
Publication year - 2020
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1576/1/012013
Subject(s) - pedestrian detection , artificial intelligence , computer science , preprocessor , computer vision , pixel , adaptability , pedestrian , algorithm , pattern recognition (psychology) , engineering , ecology , transport engineering , biology
To solve the problems existing in the traditional pedestrian detection methods, such as the complexity of pedestrians wearing clothes and the change of light leading to the decrease of detection accuracy, this paper proposes an MSER pedestrian detection algorithm combined with D-NP (Double Nine-Grid). The algorithm utilizes the characteristics of high pixels of TOF (Time Of Flight) cameras and the adaptability of MSER (Maximally Stable Extreme Regions). Firstly, the depth image is subjected to hole patching, and median filtering and morphological operations are used for preprocessing. Then the MSER algorithm combined with D-NP is used to segment the image to extract candidate head regions. Finally, We perform experiments on our dataset, and the detection recognition rate of this algorithm reaches 97.26%, which can effectively reduce the impact of light, and has a good detection performance for pedestrians carrying backpacks, hats and other complex scenes.