US 11,953,312 B2
System and method of object inspection using multispectral 3D laser scanning
Kok Weng Wong, Singapore (SG); Albert Archmawety, Singapore (SG); Jun Kang Ng, Singapore (SG); and Chee Chye Lee, Singapore (SG)
Assigned to MIT SEMICONDUCTOR (TIAN JIN) CO., LTD, Tianjin (CN)
Appl. No. 17/426,674
Filed by MIT SEMICONDUCTOR PTE LTD, Singapore (SG)
PCT Filed Feb. 1, 2019, PCT No. PCT/SG2019/050064
§ 371(c)(1), (2) Date Jul. 29, 2021,
PCT Pub. No. WO2020/159434, PCT Pub. Date Aug. 6, 2020.
Prior Publication US 2022/0107174 A1, Apr. 7, 2022
Int. Cl. G06T 15/00 (2011.01); G01B 11/25 (2006.01)
CPC G01B 11/2504 (2013.01) [G01B 11/2518 (2013.01); G06T 15/005 (2013.01)] 18 Claims
OG exemplary drawing
 
1. A system for obtaining a high-resolution 3D image of an object, comprising:
a) a first camera at a first location;
b) a second camera at a second location;
c) a third camera at a third location;
d) a red laser light source;
e) a blue laser light source;
f) a green laser light source; and
g) a processing unit,
wherein the first camera captures a first color image of the object and the processing unit separates the first color image into a first monochrome image using the red laser light source, a second monochrome image using the blue laser light source and a third monochrome image using the green laser light source,
wherein the second camera captures a second color image of the object and the processing unit separates the second color image into a fourth monochrome image using the red laser light source, a fifth monochrome image using the blue laser light source and a sixth monochrome image using the green laser light source,
wherein the third camera captures a third color image of the object and the processing unit separates the third color image into a seventh monochrome image using the red laser light source, an eighth monochrome image using the blue laser source and a ninth monochrome image using the green laser light source, wherein each camera captures three lines that are projected across the object and the three lines provide three sets of point clouds, wherein each position is scanned three times from three lasers from different angles,
wherein the processing unit converts the first monochrome image, the second monochrome image, the third monochrome image, the fourth monochrome image, the fifth monochrome image, the sixth monochrome image, the seventh monochrome image, the eighth monochrome image and the ninth monochrome image into point cloud data; and
wherein the processing unit creates a single high-resolution 3D image of the object from the point cloud data.