Skip to main content
ARS Home » Midwest Area » East Lansing, Michigan » Sugarbeet and Bean Research » Research » Publications at this Location » Publication #366281

Research Project: Nondestructive Quality Assessment and Grading of Fruits and Vegetables

Location: Sugarbeet and Bean Research

Title: Enhancing chlorophyll fluorescence imaging under sttructured illumination with automatic vignetting correction for detection of chilling injury in cucumbers

item LU, YUZHEN - US Department Of Agriculture (USDA)
item Lu, Renfu

Submitted to: Computers and Electronics in Agriculture
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 12/1/2019
Publication Date: 12/9/2019
Citation: Lu, Y., Lu, R. 2019. Enhancing chlorophyll fluorescence imaging under sttructured illumination with automatic vignetting correction for detection of chilling injury in cucumbers. Computers and Electronics in Agriculture. 168:105145.

Interpretive Summary: Cucumbers are sensitive to chilling injury when they are kept at certain low temperatures. Chilling injury first occurs below the surface of cucumbers, which are often difficult to detect visually at the initial stage. The injured tissues would deteriorate over time and finally result in the products being unsuitable for consumption. Hence, early detection of chilling injury in cucumbers is critical to ensuring quality of products for the fresh market. Chlorophyll fluorescence imaging (CFI), a technique that measures longer-wavelength light (i.e., fluorescence) released by chlorophylls in the green plant material after the excitation of ultraviolet or short-wave visible light, is known to be useful for detecting physiological disorders in green-skin products like cucumbers. In this study, we proposed a new approach by integrating the CFI technique with structured illumination, along with an image processing technique (i.e., bi-dimensional empirical mode decomposition, or BEMD) for detecting chilling injury in picking cucumbers. Two optimal wavelengths (i.e., 675 nm and 750 nm) for acquiring chlorophyll fluorescence images were first determined under uniform illumination. These wavelengths were then used in the structured-illumination chlorophyll fluorescence imaging system for detecting chilling injured pickling cucumbers. The acquired images were effectively enhanced by using the BEMD method. Results with pickling cucumbers showed that the structured illumination CFI technique resulted in significantly better images for enhanced detection chilling injury in cucumbers, compared to uniform illumination. This approach of integrating CFI with structured illumination can provide an effective means for enhanced detection of defects in horticultural products.

Technical Abstract: Chlorophyll fluorescence imaging (CFI) is useful for detecting physiological disorders or defects for green-skinned horticultural products, because defective and normal plant tissues would have different responses to ultraviolet (UV) or short-wavelength visible excitation. This study was intended to evaluate the effectiveness of a new CFI approach by using structured illumination coupled with a proposed automated method for vignetting correction of chlorophyll fluorescence images, for enhanced detection of chilling injury in cucumbers. A CFI system with UV-blue light as an excitation source under structured illumination was assembled. Spectral images over the spectral region of 660-800 nm in 5 nm increments were first acquired from chilling-treated cucumbers under uniform UV-blue illumination to determine appropriate wavebands for implementation of CFI under structured illumination. Further experiment was conducted on a larger group of chilling treated cucumbers to acquire chlorophyll fluorescence images under structured illumination for two wavebands centered at 675 nm and 750 nm. An automatic method for vignetting correction of fluorescence images was proposed by using a modified bi-dimensional empirical mode decomposition (BEMD) technique. Results showed that the chlorophyll fluorescence spectra of cucumbers were characterized by two emission peaks around the two regions of 685-690 nm and 740-745 nm. The proposed BEMD method was effective for vignetting correction of fluorescence images, which eliminates the need of using a physical fluorescence target for image correction. Moreover, structured illumination was found to provide significantly better images in terms of the image sharpness and contrast between the normal and chilling-injured tissues, compared to uniform illumination, which enhanced the detection of chilling injury in cucumbers.