Moving camera motion target detection method based on pixel modeling

• Main Authors: WAN MINJIE, GONG ZHENFEI, WANG WEIJIE, ZHAO TIEKUN, GU YUJIE, GU GUOHUA, YAO ZHEYI, WANG PENGCHENG, REN KAN, CHEN QIAN, ZHU GUOQIANG, QIAN WEIXIAN, TIAN JIE
• Format: Patent
• Language: Chinese; English
• Published: 24.10.2017
• Subjects: CALCULATING; COMPUTING; COUNTING; IMAGE DATA PROCESSING OR GENERATION, IN GENERAL; PHYSICS

The invention provides a moving camera motion target detection method based on pixel modeling. Normalization processing is carried out on a video image sequence acquired by the moving camera; normalization processing on the image can eliminate influences of lighting factors; a feature point of each frame of image in the image sequence is extracted, a feature point corresponding relationship between two adjacent frames of images is used to acquire a transformation matrix of a latter frame of image corresponding to a former frame of image, the transformation matrix is used to update a pixel model of the former frame of image and a pixel model of the latter frame of image is acquired, and the pixel model of the first frame of image in the image sequence is acquired through initialization; according to the feature point and the pixel model, decision thresholds needed for target detection and the minimum number of the decision thresholds are calculated and acquired; and according to the decision threshold and the