Brief Summary of Minutes of Annual Meeting

NE-1008: Accomplishments for period: February 2004  June 2005 1. Define and measure the physical, mechanical, optical, and other properties of fruits and vegetables and their functional relationships to quality, and establish a database of these properties. (CA, HI, NY-C, GA, ARS-MI, ME, MI, NC, NY-G, WA) USDA-MI. A hyperspectral imaging system was assembled and tested for measuring the optical properties (absorption and scattering coefficients) of turbid food materials including fruits and vegetables. Simulation turbid samples with various known optical properties similar to those of apples were created and used for evaluating the hyperspectral imaging system. Scattering and absorption coefficients were determined by fitting a diffusion theory model to the spectral scattering profiles acquired from the simulation turbid samples at wavelengths between 530 nm and 730 nm. The hyperspectral imaging system provided good measures of the scattering coefficient for all turbid samples. However, large errors were found in determining the absorption coefficient for selected turbid samples, indicating the need for further improvements in data analysis. The hyperspectral imaging technique is potentially useful for rapid, noninvasive determination of the optical properties of horticultural products and other turbid food and agricultural products. Next year. Research will be focused on three areas: 1) improving the data analysis method so that the hyperspectral imaging system can meet the requirement of measuring both scattering and absorption coefficients of turbid food samples simultaneously; 2) collecting optical properties data for selected horticultural and food products; and 3) developing simulation models to investigate and quantify light scattering and absorption patterns in intact apple fruit. GA. A wagon with springs and shock absorbers was developed for transporting peaches from orchard to packinghouse. Tests at speeds of 5, 10, 15, 20, and 25 MPH showed the shock (in vertical acceleration) from a bump was reduced by 50 to 79 percent compared with no cushioning. Tests in transporting peaches from orchard to packinghouse reduced damage only slightly because the fruit is harvested at a very firm maturity to reduce damage during handling. HI. Over 1000 spectra for coffee grown in Hawaii and other coffee coffee producing areas have been assembled and cataloged for use. Next year. Will collect taste panel, chemical data, and NIR spectra of coffee grown under three different levels of shade. Will attempt to correlate these data in predictive models. NC. The objective of this project was to develop and investigate whether a machine vision system can quantitatively describe the physical characteristics of a sweetpotato and then produce a unique number that is a more accurate descriptor of the root. NY-Geneva. We determined the effect of acid phosphatase activity on the acidity and Brix/acid ratio in apples. We found high-acid cultivars (Zappata, Cap of Liberty and Edward VII) exhibited low Brix/acid ratios while low-acid cultivars (Empire and Idared) had high Brix/acid ratios. The coefficient of correlation for acid phosphatase and Brix/acid acid ratio in apples was 0.96. An inverse relationship appeared to exist between the activity of acid phosphatase and the acidity of apples. We examined the growth and survival of beneficial lactic acid bacteria in beet juice. The four lactic acid bacteria (Lactobacillus acidophilus, L. casei, L. delbrueckii, and L. plantarum ) rapidly utilized beet juice for cell synthesis and acid production. The cultures survived the low pH and high acidity conditions in fermented beet juice even after 4 week of cold storage. Next year. We will study the effect of acid phosphatase activity on the fruit quality. We will determine the relationship between the activity of acid phosphatase and the Brix/acid ratio in cherries. We will continue our work to improve the quality and safety of fermented fruit and vegetable products. UC-Davis. The flesh color and flesh firmness of over 13,000 clingstone peaches were measured instrumentally in 2004 and compared to the current official manual inspection methods of the California Department of Food and Agriculture. The instruments evaluated were a new impact-type non-destructive firmness instrument, a universal testing instrument equipped with a Magnus-Taylor style penetrometer and a LED-based colorimeter. This study was conducted to assist the California clingstone peach industry evaluate recently developed sensor technologies for color and firmness and to assess the feasibility of replacing the current subjective methods with objective instrumental methods for determination of fruit maturity and softness. Many in the industry were unaware of the development of light emitting diode (LED) based color meters. These portable devices are much more stable than previous designs, remaining calibrated for many weeks. Also most of the industry, outside of a few peach processors, were unaware of the recent developments in non-destructive impact-type firmness sensing. This study made significant progress in introducing these technologies to the California clingstone industry and in assisting them in evaluating their potential for use in fruit inspection. Next year. Continue to investigate the application of state-of-the-art color and firmness sensing technologies in clingstone peach inspection. WA. In 2005, WA station developed a method to measure the tensile mechanical properties in fruits and vegetables from compressive material properties and bending deflection data. Next year. Measure the tensile mechanical properties of apples and pears grown in the Pacific Nortwest and attempt to correlate these values with firmness values obtained from a trained sensory panel. 2. Develop, evaluate and apply rapid non-destructive sensor technology for quantitative measurement of fruit and vegetable quality. (CA, MI, ARS-MI, PA, IN, NY-C, GA, NC, MD, WA, HI, ME) USDA-MI. Light scattering is influenced by the structural characteristics of fruit and can be used for measuring fruit firmness. A novel multispectral scattering technique was proposed and developed for measuring apple fruit firmness. In the past year, our research has been focused on improving the technique so that it can provide consistent, accurate measurement of fruit firmness. A liquid crystal tunable filter (LCTF) based multispectral imaging system was used to determine optimal wavelengths for predicting fruit firmness at wavelengths between 650 nm and 1000 nm. Spectral scattering was described by an improved mathematical model with four parameters. Model parameters were found to correlate well with fruit firmness. Improved data analysis methods were proposed for removing noise from the scattering images and for incorporating the effect of fruit size into the scattering profiles. These improvements led to significantly better predictions of apple fruit firmness (r=0.90 for both Red Delicious and Golden Delicious). A portable firmness tester, which detects force at the bioyield point of fruit as a measure of fruit firmness, was assembled and tested on apples. The tester can be operated as a standalone device or via computer with the automatic data recording and analysis functions. The tester does not degrade apple fruit after measurement and correlates well with the standard destructive Magness-Taylor firmness tester (r=0.87). Recent research showed that reflectance and fluorescence are useful for assessing postharvest quality of apples. A hyperspectral imaging system was set up for acquiring both hyperspectral reflectance and laser-induced fluorescence scattering images from apples. Algorithms were developed for extracting scattering features over the visible and short-wave near-infrared region and related to fruit quality parameters (skin and flesh color, fruit firmness, soluble solids content, and titrational acid). Reflectance generally performed better than fluorescence for predicting multiple quality parameters. Reflectance also gave excellent measures of the skin and flesh color of apples. The integration of reflectance and fluorescence improved predictions of fruit firmness, soluble solids, and titrational acid. Next year. Our research will be focused on the following three areas: 1) continuing research on multispectral scattering for assessing fruit firmness and soluble solids. We will improve light source and develop a better calibration method for measuring fruit firmness; 2) improving the hyperspectral reflectance and fluorescence imaging system for faster, more consistent measurement of fluorescence and developing an improved method of integrating reflectance and fluorescence for measuring the quality attributes of apples; and 3) developing imaging and spectroscopic techniques for detecting defects of pickling cucumbers resulting from mechanical stress, temperature injury, physiological disorder or diseases and for nondestructive measure of internal quality (firmness, dry matter, and skin and flesh color) of pickling cucumbers. GA. X-ray images were compared with visual examinations of 500 to identity diseased sweet onions during packing. The x-ray machine removed over 90 percent of the diseased onions with a false positive rate of 8%. HI. Determined that NIRS using conventional pretreatment of absorbance spectra and PLS would not yield calibration equations for DDT levels that could be validated in the 0 to 5 ppm range. The calibration set consisted of 8 sources of green bean each fortified at 0.001, 0.01, 0.1, 0.5, 1, 2, 5 ppm and subsampled 5 times for a total of 280 spectra. PLS (partial least squares) yielded equations of less than 10 factors and r^2 values > 0.7. However, predictions of DDT in samples outside of the calibration set were not within 5 ppm of GC (gas chromatograph) analysis. Next year. Will explore methods to obtain calibration model for DDT in green bean coffee. Will use kit to determine the physical properties of coffee grown in Kona. Will relate this data to design parameters for milling and storing high quality coffee. IN. During the previous reporting cycle, apples were tested using a conveyor system designed for use with a low-field (5.35 MHz) proton magnetic resonance (MR) sensor. A set of 16Rome (RO) and 25 Red Delicious (RD) apples were tested. The results, which were recently published, are as follows. There were relatively few classification errors (12.5% for RO, 0% for RD) when sorting conveyor speed was 50 mm/s. As conveyor speed increased, sorting accuracy decreased. At the fastest conveyor speed tested, 250 mm/s, the classification errors were 31.25% for RO apples and 28% for RD apples. Design features were identified that, if implemented, should increase the accuracy of sorting at the higher conveyor speeds. Next year. Results of the tests conducted on tomatoes will be summarized and submitted for publication. Preliminary tests will be conducted to evaluate the potential for using magnetic resonance for detecting internal damage or quality attributes of other types of fruit and vegetables. ME. Prediction of blueberry infestation by using a PLS model. Prediction ratio using two different NIR systems (UM-650-1050 nm, KSU-650-1580 nm) ranges between 70-80 percent. Prediction ratio includes field to field and season to season variations in the wild blueberry crop, covering three seasons (2002-2004). We have determined that the detection limit of the NIRS method based on maggot size (length) is 2 mm ~ equal to USDA standard test limit (visual, boil test). Spectral subtractions of averaged spectra from infested and non-infested blueberries show differences in the 1300-1700 nm range. In addition, comparison of subtractions of stem and calyx spectra from the same berry showed similar features for infested and non-infested berries. Next year. A. Determine which chemical compounds in the blueberry and maggot contribute to NIR signal and classification: a.1. FT-IR scan of infested and non-infested berries  preliminary determination of chemical composition, a.2. LC/MS analysis to identify functional groups and concentration ranges. B. Attempt to reduce water signal in the spectra to improve classification: b.1. Use FT-IR to calculate scaling factor for water spectrum subtraction at 2000  2500 nm, b.2. Use humidity chamber to get comparable water content in berries before scanning during 2005 field season, b.3. Remove water by freeze-drying berries and maggots and test classification. C. Determine the probability to detect signal coming directly from the maggot: c.1. Determine the probability to find the maggot close to the berry stem end using confocal microscopy and microtoming, c.2. Calculate maggot to berry volume ratio. MI. The results of previous year studies measuring reflectance and fluorescence of cherries with and without plum curculio larval infestation were used to develop a basic prototype dedicated sensor. The prototype was constructed of diodes with light outputs centered at a red and the other at a NIR wavelength. During the 2004 season the prototype sensor was evaluated in parallel with laboratory instrumentation to compare the reflectance response and classification of whole cherries and internal tissue from infested and non-infested cherries. The prototype electro-optic sensor performance as part of a tissue classification process was equal to classifications obtained through analyzing the data collected with laboratory instrumentation. Measurements with laboratory instruments were also collected in modes of transmission and interactance and evaluated against the reflectance data and compared to similar experimentation from the previous year. Measurements were collected on the common Montmorency cultivar and also on a promising newly introduced cultivar, Balaton, which is much darker in flesh color. Cherry classification accuracy ranged between 50 and 89% depending on mode of measurement and whether whole intact cherries or portions of tissue were measured. However, due to an imbalance between the number of infested and noninfested samples, the results carried large standard deviations. A study was completed and a thesis published on measuring apple tissues of varying quality under multispectral reflectance and fluorescence imaging modes. These measurements of properties were combined with neural network analysis to identify and classify five different apple disorders on three cultivars of apples. A combined eighteen fluorescence and reflectance images over a range of wavelengths were initially acquired on a large set of apples. Classification utilizing the entire (18) image data set was compared against classification using subsets of the initial set. Using optimal wavelengths, most defects could be classified with 95% or greater accuracy using three or less of the initial image set. Optimal wavebands, however, varied depending on defect and apple cultivar. Next year. The project relating to non-destructive evaluation of internal cherry quality (primarily past or present insect presence) will continue by attempting to improve on the number and distribution of various levels of cherry quality samples. Replication and verification of previous years experiments will be conducted. An arrangement has been made to evaluate samples of good versus insect damaged cherries utilizing the magnetic resonance system instrumentation at the cooperating IN NE-1008 station to test the potential of MR in differentiating the cherry tissue quality. NY. Cornell University, Zedec Technologies and the USDA/ARS Appalachian Fruit Research Station cooperative efforts led to the development of a prototype inspection system for processing apples. High speed handling/orienting systems were developed. Surface defects were quantified in terms of apple volume to allow USDA grades for processing apples to be determined. This system will be available at USDA/ARS Appalachian Fruit Research Station for inspection and testing. A summary of the results of this effort has been published. Minimal effort has been continued with Zedec Technologies on development of high speed handling/orienting system for Xray inspection of apples. A higher intensity X-ray source is needed for this work. PA. Research continued to combine the surface-acoustic wave-based "zNose" technology with the Cyranose 320 through sensor fusion to detect damage in apples. The response of the e-nose 32 sensors are being combined with the compound-specific peaks from the zNose through artificial neural networks or multivariate analysis techniques to determine presence of compounds associated with damaged apples. Tests using the individual instruments showed that both the e-nose and the Znose could distinguish healthy from damaged apples (those with purposely inflicted cuts) after about 5-7 days. The successful classification rates of good vs. bad apples ranged between 85 and 100%. The goal is to use sensor fusion to allow for earlier and more accurate detection of damage. Direct differentiation and quantification of microorganisms in apple juice was accomplished by FTIR spectroscopy. FTIR absorbance spectra of ten microorganisms in an apple juice matrix were used with a Probabilistic Neural Networks (PNNs) based algorithm for differentiation and quantification. The absorbance spectra in the fingerprint region 600-1750 and 2800-3200 and principal components of the spectra in the 600-4000 cm-1 region were used separately in the analysis. Although no significant difference in the results were observed the time taken for data processing and training of the PCA based PNNs (PCA-PNNs) was reduced by about 30-45 minutes compared with the data from the fingerprint area, which took about 70-90 minutes. The prediction accuracy using the two methods was 93% for differentiation and 90% for quantification. UC-Davis. A rapid, quantitative biosensor technique for fungal contamination in raw and processed tomato juice was refined in 2004 and adapted for outdoor use at processing tomato inspection stations. The main accomplishment was to adapt the UC Davis mold assay for use by seasonal inspectors in an uncontrolled environment of a tomato inspection station. Existing blood cell washing technology was adapted for use with the mold biosensor method, allowing automation of the washing step of the assay. 3. Develop, evaluate, and apply rapid sensing technologies to assure food safety including bio-security, purity, and integrity of produce. (CA, PA, MI, NY-C, GA, WA) GA. Sweet onions were sampled weekly at three retailers. Size, mass, decay, defects and other quality attributes were determined according to the USDA grade standards. The mass of bagged onions typically exceeded labeled amounts. Differences were noted in onion quality available to customers. The amount of decay and sunscald differed by grower/shipper and during the season. The quality of the loose onions was more variable than for bagged onions. HI. Parallel rollers were shown to effectively separate coffee pulp from coffee beans in hulls. A series of experiments determined the optimal set of conditions for the separation. At these conditions 100% of pulp is expelled while all beans are retained without any damage to the hull of the beans. An innovative processing system is in development around this technology. Next year. Will build and test a prototype separator based on the roller technology. Industry (SI Seiko Co., Ltd.) Fruit grading system and grading robot with database for traceability. Next year. 1. Development of a new type grading machine and 2. Construction of fruit information center to accumulate data from the grading system and soil analysis system for farming guidance to producers and for traceability information to consumers and distributors. MI. We developed several biosensors designs during the reporting period: two antibody-based and two DNA-based biosensors. These biosensors were tested in pure culture of E. coli O157:H7, E. coli, Salmonella spp., and Bacillus spp. The antibody-based biosensors were also tested on selected fruits and vegetable samples. Next year. Further develop and validate the biosensors in fresh produce and other food products. NY. A new sensing technology that makes use of DNA is under development (Nature biotechnology, published on-line 12 June 2005). DNA can be synthesized to form a Y. PA. Research continued on the application of surface plasmon resonance biosensors to detect pathogenic microorganisms, including Escherichia coli 1O57:H7 and Staphylococcus aureus. We are currently studying the application of array-based biosensors using SPR imaging to detect antibody/antigen interactions with the goal of simultaneous detection of multiple pathogenic bacteria. SPR Imaging is an extension of the traditional SPR devices that are well known for their high sensitivity, real-time monitoring, and label-free detection of probe-target interactions. By replacing the light source and optical sensor that would be found in a traditional SPR unit with a collimated white light and an imaging CCD, the SPR phenomena of the entire surface of the sensor can be monitored. SPR imaging offers improvements over traditional SPR because it allows for multiple probe/target interactions to be monitored simultaneously. UC-Davis. SYBR Green Assay. A three-step SYBR Green real-time PCR assay using the Sen primer set (Csordas, Barak, and Delwiche, 2004) and an annealing temperature of 57°C was used to detect Salmonella enterica (S. enterica) extracted DNA. A flow-through PCR system was tested with S. enterica Newport extracted DNA at concentrations of 100, 10, and 1 ng per 50 µl. Additionally a concentration of 1000 ng of extracted S. enterica Newport DNA per 40 µl was tested. A total PCR mix volume of 55 µl was used for the 100, 10, and 1 ng samples and a 40 µl reaction volume was used for the 1000 ng sample. New silicone and glass reaction chamber tubes were used for each PCR. Next year. We are planning on working with the cell washing equipment manufacturer to develop an adaptable version of their cell washer that will allow control of both the binding step and the washing step of the bioassay for mold. WA. We have studied the influence of RF based pest control treatments on quality of oranges and apples. Next year. Research to reduce heating time and minimize off-flavor development.