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 regions of 685–690 nm and 740–745 nm respectively. 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, compared to uniform illumination, structured illumination was found to provide significantly better fluorescence images in terms of the image sharpness and contrast between the normal and chilling-injury tissues, which were inductive to enhancing the detection of chilling injury in cucumbers.