Early Detection of Mold-Contaminated Maize Kernels Based on Optical Coherence Tomography

Abstract

Fungal infection is a pre-harvest and post-harvest crisis for maize kernels and infection detection at early stage is still a difficult point. The feasibility of optical coherence tomography (OCT) for imaging mold-contamination on maize kernel at early stage is investigated. OCT technique provides high-resolution 2D tomographic images of micro-structure of maize and realizes immediate, automated, and non-destructive detection of mold by proposed method. After denoising, edge detection, smoothing, and flatten, thin pericarp layer is cropped and extracted. Then, the number of non-zero pixels on the upper boundary is calculated and the float threshold is used to classify whether there are mildew pixels in a certain column. All the experimental results suggest that mold contaminated maize kernel can be detected and marked by monitoring near surface layers combined with our proposed method. The OCT approach is suitable for fast, nondestructive diagnosis of mold at early stage, which may potentially lead to significant cost savings and better control of the spread of mold in the food industry.