Submitted to: Michigan State University Cucumber Reporting Session
Publication Type: Experiment Station
Publication Acceptance Date: 1/10/2010
Publication Date: N/A
Citation: N/A Interpretive Summary:
Technical Abstract: This report briefly summarizes the three studies performed in 2009 on cucumber defect detection. Hyperspectral imaging transmittance mode has demonstrated great potential for online sorting and grading of cucumbers and pickles. However, the technique is still limited by its speed in acquiring and processing spectral images. Therefore, we investigated a different lighting design in order to improve the speed of acquiring hyperspectral images from cucumbers. In another study, laser scattering technique was explored as a simpler means for detecting cucumber defect. In addition, an experiment was performed on using computer tomography (CT) to obtain 3-dimensional images of the internal structure of pickling cucumbers for defect detection. Preliminary analysis showed that the new illumination design using near-infrared light-emitting diodes (LEDs) with the wavebands of 830 and 890 nm greatly improved the image acquisition speed (2 ms), while reducing cost and energy use. Up to 90.3% accuracy for defect classification was achieved. The laser scattering study showed that the position of laser relative to the imaging device had a significant effect on defect detection; the zero degree position was the most effective for defect detection, achieving the maximum accuracy of 100%. The technique is simpler and easier to implement, and hence it has great potential for internal defect detection. CT scan images could effectively identify hollow or split center in the cucumbers. However, the technique was not effective for detecting bruises caused by mechanical stress. While CT scan imaging is potentially useful for inspecting internal quality of cucumbers, it is still expensive at the present.