MAIZE KERNEL ANALYSIS
About the sample: Maize
- In this example, a fully grown ear of maize was scanned using a gantry geometry micro-CT scanner, CT Lab GX.
- The CT image was segmented into five phases using the deep learning segmentation technique. Then, watershed transform was used for object separation of maize kernels.
- The volume distribution was calculated for the maize kernels.
1. CT scan
An ear of fully grown maize was scanned to produce the 3D grayscale CT image. A 3D cross-section shows kernels, outside silk and husk as well as the internal pith structure.
2. Image segmentation
Deep learning segmentation was used to segment the maize ear into a husk, kernels, pith, sample holder, and air. Then, kernels were separated using the watershed transform method.
3. Kernel size analysis
Kernel volumes were calculated to allow a volume distribution analysis. The average kernel volume was 96.97 mm3. A volume distribution histogram with a color-coding overlay on the separated kernels is shown.