公开(公告)号：US20180075605A1

公开(公告)日：2018-03-15

申请号：US15263761

申请日：2016-09-13

Applicant: Intelligent Fusion Technology, Inc

Inventor： BIN JIA ,  KUI LIU ,  SIXIAO WEI ,  ERIK BLASCH ,  CAROLYN SHEAFF ,  HAIBIN LING ,  GENSHE CHEN

IPC: G06T7/11 ,  G06T7/194

CPC classification number: G06K9/4642 ,  G06K9/0063 ,  G06K9/00986 ,  G06K9/38 ,  G06K9/4671 ,  G06K9/6211 ,  G06K9/6269 ,  G06T3/4038 ,  G06T7/33 ,  G06T2207/10032 ,  G06T2207/30184 ,  G06T2207/30236

Abstract: In accordance with various embodiments of the disclosed subject matter, a method and a system for detecting multiple objects from real-time images are provided. The method comprises: performing, using a CPU host, an image segmentation process to divide real-time input images into a plurality of image partitions; performing, by multiple GPUs, a fast block-wise registration process, a mark setting process, a background generation process, a foreground generation process based on a Hyper-Q computation infrastructure, and a support vector machine classification process; and generating, by the CPU host, visualization classification images.

公开(公告)号：US20170124758A1

公开(公告)日：2017-05-04

申请号：US14927428

申请日：2015-10-29

Applicant: Intelligent Fusion Technology, Inc

Inventor： BIN JIA ,  KUI LIU ,  GENSHE CHEN ,  KHANH D. PHAM ,  ERIK BLASCH

IPC: G06T17/00 ,  G06T19/00

CPC classification number: G06T17/00 ,  G06F9/46 ,  G06F17/5009 ,  G06F2217/16 ,  G06T19/006 ,  G06T2210/21

Abstract: Method and system for predicting collision probability of space objects are provided. Space objects are allocated according to a total number thereof for parallel computation in a GPU device. Initial orbit states of the space objects are generated by an initial-orbit-state generation kernel of the GPU device. Orbit propagation for each space object is performed based on the initial orbit state to predict a state of each space object by an orbit propagation kernel of the GPU device. The predicted states of the space objects are collected by a predicted-orbit-state collection kernel of the GPU device. A real-time visualization and rendering of the predicted space objects are performed in a form of point cloud during interactions between CPU threads and GPU kernels. A CPU host predicts a collision probability of any two space objects among the multitude of space objects, according to the interactions between the CPU threads and the GPU kernels.