公开(公告)号：US11922666B2

公开(公告)日：2024-03-05

申请号：US17228731

申请日：2021-04-13

Applicant: PixArt Imaging Inc.

Inventor： Keen-Hun Leong ,  Kheng Jin Ooi ,  Yueh Mei Kim

IPC: G06V10/00 ,  G06V10/145 ,  G06V10/50 ,  G06V10/75

CPC classification number: G06V10/145 ,  G06V10/507 ,  G06V10/751

Abstract: An object detection system and method includes: an optical image sensor arranged to perform the following steps: capturing a calibration image during a calibration stage, dividing the calibration image into a plurality of quadrants, and calculating a parameter for each of the quadrants; capturing a plurality of raw images during a detection stage, dividing each image of the raw images into a plurality of quadrants, and calculating a parameter for each of the quadrants; comparing the respective parameters of each quadrant of a raw image with the respective parameters of each quadrant of the calibration image to generate a ratio value for each quadrant; and comparing the ratio value of each quadrant with a predetermined threshold. When each ratio value of specific quadrants of the quadrants is greater than the predetermined threshold, object detection is confirmed.

公开(公告)号：US20220327325A1

公开(公告)日：2022-10-13

申请号：US17228731

申请日：2021-04-13

Applicant: PixArt Imaging Inc.

Inventor： KEEN-HUN LEONG ,  Kheng Jin Ooi ,  Yueh Mei Kim

IPC: G06K9/62 ,  G06K9/46 ,  G06K9/20

Abstract: An object detection system and method includes: an optical image sensor arranged to perform the following steps: capturing a calibration image during a calibration stage, dividing the calibration image into a plurality of quadrants, and calculating a parameter for each of the quadrants; capturing a plurality of raw images during a detection stage, dividing each image of the raw images into a plurality of quadrants, and calculating a parameter for each of the quadrants; comparing the respective parameters of each quadrant of a raw image with the respective parameters of each quadrant of the calibration image to generate a ratio value for each quadrant; and comparing the ratio value of each quadrant with a predetermined threshold. When each ratio value of specific quadrants of the quadrants is greater than the predetermined threshold, object detection is confirmed.