Wire Rope Dataset

Citation Author(s):
Kuosheng
Jiang
Submitted by:
kuosheng jiang
Last updated:
Wed, 02/08/2023 - 07:34
DOI:
10.21227/dm5z-2013
Data Format:
License:
0
0 ratings - Please login to submit your rating.

Abstract 

  We use industrial cameras to take images of steel wire ropes under different conditions, and use these wire rope images to train the U_Net network, and realize the semantic segmentation of the wire rope images by the U_Net network.

 

Instructions: 

 We use industrial cameras to take images of steel wire ropes under different conditions. We put the images of steel wire ropes in five folders, named as:Camera position step up_1; Camera position step up_2; From dark to light; Rotate(360 degrees); Rotate(360 degrees). Images in different folders come from different sources, explained below: Camera position step up_1:Move the camera from bottom to top to obtain images of different positions of the wire rope.  Camera position step up_2: The camera rotates at a certain angle with the wire rope as the axis and then moves the camera from bottom to top to obtain images of different positions of the wire rope.  From dark to light:Adjust the brightness of the light source to obtain the images of the wire rope under different brightness. Rotate(360 degrees): Rotate the wire rope 360 degrees and randomly take images of the wire rope at different angles. Rotation(free):Apply a certain torque to both ends of the wire rope and then suddenly remove the torque applied to both ends of the wire rope, and randomly take images during the rotation of the wire rope. In addition, the dataset also provides the json file generated manually using labelme. Note: If the network model fails to be trained using the json file, you can consider converting the Chinese in the json file to English. Finally, we also provide dataset usage instructions in the wire rope dataset folder.