"Polarized Bamboo Chopsticks Dataset"

"This dataset contains polarization images of bamboo chopsticks collected using a polarization imaging camera installed on an industrial simulation platform designed to replicate real production line conditions. The images capture various surface defects that may occur during bamboo chopstick manufacturing and circulation processes, such as cracks, notches, and discoloration. Compared with conventional RGB imaging, polarization imaging effectively suppresses interference caused by natural bamboo textures and specular reflections, thereby enhancing defect contrast and improving visual separability between defective and normal regions. All images were manually annotated using the LabelImg annotation tool, where bounding boxes were assigned to defective regions. The dataset is intended for machine vision research and supports the development, training, and evaluation of automated defect detection algorithms based on deep learning frameworks"