Location: Sugarbeet and Bean ResearchTitle: Fast demodulation of pattern images by spiral phase transform in structured-illumination reflectance imaging for detection of bruises in apples
|LU, YUZHEN - Michigan State University|
|LI, RICHARD - Michigan State University|
Submitted to: Computers and Electronics in Agriculture
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 5/1/2016
Publication Date: 7/13/2016
Citation: Lu, Y., Li, R., Lu, R. 2016. Fast demodulation of pattern images by spiral phase transform in structured-illumination reflectance imaging for detection of bruises in apples. Computers and Electronics in Agriculture. 127:652-658.
Interpretive Summary: Currently, computer imaging technology is widely used for quality inspection of horticultural and food products. Uniform, diffuse illumination is the standard in most imaging applications, as it is simple and easy to implement. But conventional diffuse illumination is not so effective for detecting certain subtle features or defects underneath the skin of products (e.g., bruises in apples). A new structured-illumination imaging reflectance (SIRI) system was recently developed in our laboratory, and it was able to detect fresh bruises in apples, which otherwise could not be achieved using uniform illumination. However, the implementation of SIRI requires acquisition and subsequent demodulation of three images for a product subjected to a sinusoidal pattern of illumination, with each being shifted by a specific phase angle. It needs a long image acquisition time with accurate phase shifts. This research was therefore aimed at developing a new method of demodulating images that only needs two or one image, in order to improve the implementation of SIRI technique in terms of speed and performance. Three demodulation methods, i.e., two-phase, two-image, and one-phase spiral phase transform (SPT), were proposed and compared with conventional three-phase demodulation methods. Simulations and experiments on detecting bruises in apples were conducted to evaluate the effectiveness of the proposed SPT methods. It was found that the two-phase SPT method, which only needs two images with arbitrary phase shifts, achieved the same or better performance, compared with conventional three-phase method. The one-phase SPT method could also be used for image demodulation, although its performance still needs improvement. Spiral phase transform provides an effective means for demodulating pattern images, while reducing the image acquisition time by one third, which is a significant step towards the goal of implementing SIRI technique for real-time quality evaluation of horticultural and food products.
Technical Abstract: A structured-illumination reflectance imaging (SIRI) system was recently developed in our laboratory for enhanced quality evaluation of horticultural products. It was implemented using a digital camera to acquire reflectance images from food products subjected to sinusoidal patterns (or other similar patterns) of illumination, as compared to conventional diffuse or uniform illumination. The reconstruction of amplitude images is a key step in implementing the SIRI technique. Conventional methods require acquisition and subsequent demodulation of three phase-shifted pattern images, which limits the speed of image acquisition. This study proposed the use of spiral phase transform (SPT) for demodulation of pattern images acquired by SIRI. Three SPT-based methods involving only one or two images were investigated through numerical simulation, followed with experiments on the detection of fresh bruises in apples. Compared to conventional three-phase based and two-phase based Gram-Schmidt (GS) orthonormalization demodulation methods, SPT achieved the same or even higher demodulation accuracy when two pattern images with arbitrarily shifted phases were used. SPT also allowed single-image based demodulation, although its performance still needs to be improved. This study demonstrated that SPT for amplitude demodulation can increase the imaging speed of SIRI by using one or two pattern images, which is a significant step towards real-time implementation of the technique.