Method for detecting obstacle of rubber tired crane at container port based on binocular vision

Abstract

The invention discloses a method for detecting an obstacle of a rubber tired crane at a container port based on binocular vision. The method comprises the following steps: step 1, collecting left andright images of an installed binocular camera; step 2, respectively correcting the left and right images, and removing distortion; step 3, by using the left image as a reference image, detecting a road line (region of interest (ROI)) on the image; step 4, enabling an obstacle detection module to detect the obstacle; step 5, judging the blocking of the camera; step 6, simply tracking the obstacle;if the obstacle moves towards the inner side of the road line (ROI), alarming.

Description

technical field [0001] The invention relates to a binocular vision-based obstacle detection method for a container terminal bridge. Background technique [0002] Rubber Tire Gantry (RTG, Rubber Tire Gantry) and Rail Mounted Gantry (RMG) (collectively referred to as yard bridges), as large-scale machinery in ports, play a vital role in port operations. Its efficiency, safety, and correctness of operation have an important impact on terminal operations. The yard bridge is mainly responsible for the container transfer between the container and the container stacked in the yard and the container horizontal transportation equipment (container truck or automatic guided transport vehicle AGV, running on the truck lane), which is characterized by complex operating environment and high risk factor , the driver has poor vision and is highly dependent on the driver. [0003] During yard and bridge operations, due to the high position of the driver's cab, the unsatisfactory lighting c...

AI Technical Summary

Problems solved by technology

[0005] 1. LiDAR is expensive

[0006] The cost of laser radar equipment is very high. A 32-line laser radar costs about 20,000 US dollars, while a 64-line laser radar costs as much as 80,000 US dollars.

[0007] 2. Lane lines cannot be detected without modification

[0008] The lidar-based obstacle detection system relies on reflective markings. If no special paint is applied to the lane line, the lidar-based detection system cannot detect lane lines and other signs, and can only manually delineate the detection area, or in the lane line Special coatings are applied on them, which brings a large labor cost to different ports

At the same time, it is difficult to maintain after installing reflective markers, which also increases the cost

[0009] 3. It is difficult to classify and identify obstacles

[0013] Monocular vision obstacle detection If the vision system wants to calculate the obstacle distance, it can only install the camera at a fixed height, direction, etc., and the deviation is large, or the distance of the obstacle can only be estimated, not accurately calculated , there is also a certain deviation

[0014] 2. There will be misrecognition of flat images

[0017] If the obstacle is very close to the camera, the system needs to alarm at this time, and the monocular visual obstacle detection system often cannot work normally in such a very dangerous situation, and there are serious safety hazards

[0018] 4. Only specific types of obstacles can be detected

[0020] Unable to detect obstacles other than training data

Images