Watersheds in Digital Spaces: An Efficient Algorithm Based on Immersion Simulations
IEEE Transactions on Pattern Analysis and Machine Intelligence
Inconstancy of finite and infinite sequences
Theoretical Computer Science
Machine Vision and Applications
Hi-index | 0.00 |
An image analysis procedure was developed to quantifymorphological characteristics of convolutions inindividual cotton fibers without pre-tensioning ororientation requirements. The image of each fiber wascaptured by a PC-based color imaging system using aconventional microscope. Ends of individual cottonfibers were glued on a microscope slide without anytension or straightening. A modified watershedtechnique was implemented to identify individualconvolution segments, which were defined as sectionsof the fiber bordered by two neighboring convolutions. Length, area and perimeter of each convolution segmentwere measured directly from the image. Average width,shape factor and number of convolution segments in mmwere calculated from the measured parameters. Performance of the image analysis algorithm wascompared with visual inspection for number andposition of convolution segments in three differentvarieties of cotton. Image analysis results agreedwith visual inspection in 89.6% of the testedimages.