US 12,325,981 B2
System and method of controlling construction machinery
Cavin Lee, Yongin-si (KR)
Assigned to HD HYUNDAI INFRACORE CO., LTD., Incheon (KR)
Filed by HD HYUNDAI INFRACORE CO. LTD., Incheon (KR)
Filed on Jan. 13, 2022, as Appl. No. 17/575,299.
Claims priority of application No. 10-2021-0005089 (KR), filed on Jan. 14, 2021.
Prior Publication US 2022/0220707 A1, Jul. 14, 2022
Int. Cl. E02F 9/26 (2006.01)
CPC E02F 9/264 (2013.01) [E02F 9/261 (2013.01)] 11 Claims
OG exemplary drawing
 
1. A control system for construction machinery, the control system comprising:
a camera installed in a work apparatus to photograph a working area in which the work apparatus works;
an image processing device configured to recognize a shape of an attachment of the work apparatus in an image captured by the camera and perform a tracking image process on the image such that the attachment is displayed in a central region of the image; and
a display device configured to display the tracking image-processed image,
wherein during the tracking image process, the image processing device is further configured to calculate a relative distance from a pixel position of the attachment of the work apparatus in the image to the central region, and to move a display position of the attachment of the work apparatus to the central region based on the relative distance,
wherein the image processing device is further configured to adjust a size of the attachment in the image according to a preset tracking image processing condition to match a resolution of the display device, to thereby resolve a visual distortion caused by a size difference due to a movement trajectory of the attachment, and
wherein the image processing device includes:
a shape recognizer configured to compare the actual image of the attachment in the image with a learning image of the attachment that is recognized and stored in advance by machine learning, and
a storage portion configured to store the learning image of the attachment by executing a deep learning algorithm using the actual image received from the shape recognizer as input data.