Optical coherence tomography (OCT) is a high-resolution, non-destructive imaging modality gaining attention in agricultural research. With growing concerns about harmful chemicals like calcium carbide and acetylene gas used for artificial fruit ripening, especially in India, rapid assessment techniques for fruit health are needed. In this work, we used OCT to examine microstructures in the sub peel region of mango (Mangifera indica) ripened artificially (with ethylene gas) and naturally (in paddy straw). Qualitative analysis of OCT images showed distinct sub-peel microstructural changes. In ethylene based ripened mangoes, randomly shaped void spaces resembling intercellular spaces were formed in the parenchyma cell region. This is due to the early cell wall disintegration indicating the faster tissue ripening. Also, irregular spacing (∆t) between two distinct regions (exocarp and mesocarp) was evident as early as day 4. Conversely, naturally ripened mangoes exhibited only few irregularly shaped micro crack-like features in the outer mesocarp region only by day 6, reflecting a more gradual transformation.

A spatial resolution enhancement technique leveraging smoothed pseudo Wigner-Ville distribution (SPWVD) for Spectral Domain Optical Coherence Tomography (SDOCT) is proposed and investigated in this work. SPWVD, commonly used in communications for time-frequency analysis, is adapted here to exploit the apparent non-stationary behavior of SDOCT interferometric signals. This behavior in SDOCT refers to specific frequencies in the signal that remain unresolved when analyzed globally, but become resolved when examined locally using time-frequency tools. Conventional SDOCT utilizes a Fast Fourier Transform (FFT) to extract the global frequency content of the signal. The process can potentially miss closely spaced frequencies. In addition, using the FFT method to generate axial scans (A-scan) compromises axial resolution due to the complex conjugate artifact, which uses the N-pixel detector data sub-optimally. The proposed technique analyzes the local spectral characteristics of an N point interferometric signal and generates high resolution A-scans over N samples rather than N/2, thereby overcoming some limitations of the FFT method. The experimental results demonstrate a twofold increase in A-scan resolution, effectively resolving closely spaced layers in the sample. The B-scan images processed with the proposed method also show improved contrast to noise ratio compared to the FFT method enabling high resolution OCT imaging.

This work reports the application of smoothed pseudo Wigner-Ville distribution for optical signal processing in spectral domain optical coherence tomography. Although traditional Fourier Transform methods have been successful in generating Axial scans, they are limited in fully exploiting the resolution of data captured by N-pixel detectors due to the complex conjugate symmetry property, resulting in redundant information over N/2 number of the pixels. To overcome this limitation, our proposed approach leverages smoothed pseudo Wigner-Ville distribution for precise frequency analysis across N samples and incorporates Kaiser windows to achieve effective frequency smoothing. This technique enables the generation of Axial scans over N pixels instead of N/2 pixels. Experimental verification validates the efficacy of the proposed technique, demonstrating a twofold improvement in longitudinal spatial resolution compared to the conventional Fourier based method.

Continuous wavelet transformations are used in the current simulation model for the spectral domain OCT image reconstruction. This filter bank approach enhances longitudinal spatial resolution by a factor of two over the Fourier-based reconstruction technique.