WSU Tree Fruit Research & Extension Center

Postharvest Information Network

Sunday, July 24, 2016

WSU-TFREC/Postharvest Information Network/Near Infrared Sorting for the Washington Apple Industry



Near Infrared Sorting for the Washington Apple Industry


Introduction

My introduction to Near Infrared sorting of apples was when Walt Hough of Auvil Fruit and I were requested by the Washington State Tree Fruit Research Commission to travel to Japan in December 1995.

The goal of the trip was to evaluate the potential for employing a newly developed apple and peach sorting system, using Near Infrared (NIR), on packinglines in Washington State. A secondary goal was to investigate how Japanese growers determine the maturity of Fuji apples and how they are stored and packed (this will be covered elsewhere).


NIR Sensing of Apples in Japan

Travel took place during the week of December 17-22, 1995. Three full days were spent working in Japan. We were accompanied by a translator, Mrs. Kay Nakagawa, and her husband, Glenn, without whom the trip would have been extremely difficult.

Background on Near Infrared Technology
Near Infrared light (NIR) was first used on food crops by Dr. Norris at USDA in Beltsville on grain in 1970s. Dr. Gerald Dull (at University of Georgia, Athens) published a number of papers on the use of NIR to sort fruits and vegetables as early as 1984. Since then a number of other American (including Bruce Upchurch and Michael Delwiche) and foreign (Dr. Kawano, Japan; Drs. Schaare and Jordan et al., New Zealand) researchers have explored this technology.

Scientists have investigated the use of NIR to measure:

  • Soluble solids in cantaloupe
  • Dry matter in potatoes
  • Protein in wheat
  • Bruise detection in apples
  • Soluble solids in peaches
  • Defect sorting in stone fruits
  • Individual sugars in dry mixtures
  • Maturity of papayas
  • Sugar content of fruit juices
  • Total leaf nitrogen in tree fruit canopy leaves, etc.

NIR can be used either in the reflectance mode (bouncing off or absorbed by the fruit) or in the transmission mode (passing through the fruit).

NIR wavelengths (800-2500 nm) are just slightly longer than those of visible light (400-750 nm) and are considered part of the infrared portion of the spectrum. Infrared (IR) can be used to measure temperature nondestructively.

NIR energy is absorbed by certain chemical groups (i.e., CH, OH, and NH) and not by others, allowing for detection and quantification of certain compounds. Among these compounds are chlorophyll, soluble solids, proteins, etc.

This is not as easy as it appears, as many compounds have certain chemical groups in common; thus, the challenge becomes to distinguish one compound from another. Luckily, each compound has certain harmonics or overtones specific to it. These overtones can be analyzed allowing for identification and quantification. The use of NIR has required complex multiple linear regressions which require powerful computers. In the early stages, the development of the use of NIR was limited by the computing power.

The measurement of fruit firmness is very difficult since many factors are involved. It is proposed that NIR might bring us closer to measuring firmness, through the evaluation of "other factors." It is theoretically possible that the scattering of light within the fruit may relate to firmness.

The measurement of starch with NIR is difficult since the concentration of starch is so low, making quantification very hard. Starch could be detected by transmission between 800-1100 nm, but both detection and quantification are difficult.

A source of scientific information on NIR can be found in the book Practical NIR Spectroscopy, by B.G. Osborne, F. Fern, and P. Hindle, published by Wiley and Sons, NY, in 1993, ISBN #0-582-09946-3.

In a nutshell, NIR as currently used commercially is the use of a light source shining on or through a fruit. Some of the light is absorbed and some is reflected or transmitted through the fruit, depending upon its application.

The Limitations of NIR
The limitations for the use of NIR are both the chemical constituents to be measured and the physical problems associated with light reflectance and transmission through the fruit.

NIR can be used to detect only chemical compounds which contain CH, OH, or NH groups. Identification and quantification of the compound must be worked out by a computer. As fruit is mainly water, differentiation between water and chemical constituents of interest can be difficult.

The orientation of the fruit to the light beam may also be a problem.

NIR Sorting in Japan
Apparently, there are three companies which have, or soon will have, employed NIR in the fruit industry. The primary goal has been to develop machines capable of measuring soluble solids (SS) as fruit pass over a commercial packingline. This has been accomplished for a number of crops. A new goal has been to develop a machine which would determine acidity at the same time.

There are estimates of 60 NIR units in use in Japan in 1996. The first units were developed for peach packinglines and tested in 1989. This has spread to the apple industry in the last few years.

A new unit is to be commercially tested which is reputed to also measure acidity. This new unit is the result of a 7-year effort of a government/private enterprise partnership on behalf of the mandarin orange industry, which is a huge industry in Japan. The measurement of acidity is difficult as it requires the measurement and interpretation of transmitted rather than reflected light.

How Does Commercial Equipment Work?
Currently, apples are placed on either a belt or on individual cups by hand with the stem end up. The apple passes through a shoebox-sized instrument which measures temperature using IR. It then passes through another similar box which shoots NIR light at it. The reflected light is read by a sensor and level of soluble solids is computed from the amount of light absorbed at certain wavelengths in the NIR range vs. the amount reflected. In the next step, the apple passes through a larger box containing a skin color sensor and either a volumetric or weight sizer.

How Is the Equipment Being Used in Japan?
NIR is currently being used to sort peaches and apples by soluble solids level. We visited two packinghouses in Aomori Prefecture where NIR is used on apples.

At this point, NIR is used to provide the grower with information on SS so that he or she may change practices to optimize SS. In the wholesale market, we saw a number of boxes of fruit marked that had been subjected to NIR evaluation. We were told this fruit could be sold for higher prices. The use and market acceptance of NIR sorting on stone fruits is greater than on apples, possibly as a result of the length of time it has been used in that industry.

Mitsui Company has formulated a "ripeness index" from SS and skin chlorophyll numbers. In no case did we see boxes in the wholesale market on which actual SS or "ripeness index" numbers were written--only the word "sensored." Several people in the market reported that having the term "sensored" was helpful to sales.

Mitsui Company Research and Development
Mitsui Company hosted much of our trip to evaluate NIR. Mitsui Mining is a huge Japanese mining company which has used NIR for years to search the earth for minerals, especially quartz and calcite. In 1986, they tested the ability of NIR reflectance to measure SS in peaches. In 1989, they introduced instrumentation to the peach industry in Japan. The same year, they began to work on apples and, in 1992, they introduced the first NIR sensor for apples.

The SS relationship measured by NIR is linear between 0 and 20%. Two NIR sources are utilized with a detector between them. Accuracy of detection is highest at 3 fruit per second, but it is possible to measure SS at 6 fruit per second, but with reduced accuracy.

What Is NIR Being Used to Measure in Japan?
As described above, NIR can be used to quantify a number of different compounds. The use of NIR in the reflectance mode has been commercialized. Studies are underway to explore the use in the transmission mode.

Discussions with Mitsui representatives and people in the wholesale market indicated to us that, at the present time, SS measurement was the most useful and reliable. The measurement of skin chlorophyll appears to be used, but to a lesser degree.

Flesh chlorophyll measurement may be possible. However, in the reflectance mode, NIR penetrates only up to 5 mm into the fruit so the amount of flesh sampled is very small.

The measurement of acidity is being developed; however, this is dependent upon perfecting an NIR transmission system.

Some discussion was held regarding the ability to nondestructively measure watercore, firmness, and internal and external blemishes. These are a number of years from being available. Of these, watercore appears the most feasible.

One fruit marketer mentioned, "In the future, fruit will not look as good as it does now, since orchardists will not use bags or pluck leaves. We will use the SS sensor to assure consumers of good flavor in spite of lesser appearance."


More Recent NIR Research

In February 1997, I had an opportunity, again funded by the Washington Tree Fruit Research Commission, to attend an international symposium on sensors for nondestructive testing. The following research is reported in the proceedings from the Sensors for Nondestructive Testing International Conference and Tour, Orlando, Florida, February 18-21, 1997: "Sensors for Nondestructive Testing: Measuring the Quality of Fresh Fruits and Vegetables," available from Northeast Regional Agricultural Engineering Service, Cooperative Extension, 152 Riley-Robb Hall, Ithaca, NY 14853-5701. Here is some of what I learned about NIR.

  • In Belgium, working on apples, E. Moons et al. developed an optical fiber system using NIR to determine soluble solids. Their system runs at 3 apples per second. They have had some success measuring acidity and hardness with the peel intact. They are currently working on ways to decrease the cost of the equipment while increasing speed.

  • Work on NIR has continued at the University of Georgia on determination of SS content of peach fruits. This work is being done by K.H.S. Peiris et al. of the Department of Horticulture. Their results were very positive; however, orientation of the fruit and differences in the SS content within the flesh are obstacles which must be overcome prior to using the technology on a commercial scale. It was interesting that the SS for one variety of peaches over several seasons ranged from 6.8 to 21.2%. There is not this much variability in apples.

  • NIR is used to analyze dry matter and sugars in whole potatoes. The authors, M.G. Scanlon et al. of the University of Manitoba, point out that although NIR has been used to analyze dry matter in cereal products, its use in products containing high moisture has been difficult due to the obscuring effect of the water. They attempted to use NIR to dictate how to process potatoes into frozen french fries--a decision based on dry matter content. Of course, one of the first problems was to decide where on the potato to sample, as sugar content varies within the tuber. Their results were very promising.

  • NIR is being investigated as a method of predicting eating quality of kiwi fruit by an engineering group in New Zealand. Dr. R. B. Jordan et al. used a method they call "interactance" where a broad spectrum of light is passed through regions local to the light source and detector. This is different than transmission (passing through the fruit) or reflectance (bouncing off the fruit's surface). They use a model which assumes that the final soluble solids is equivalent to the dry matter minus "other constituents." The other constituents are reasonably constant, thus SS = final eating quality. Kiwi fruits contain a number of different sugars which determine their complex taste. Their conclusion was that, although NIR could determine edible quality of the kiwi fruit sampled, analysis of the data obtained from the NIR data would be difficult and currently is far from commercially viable.

  • Sugar determines harvest date of fresh dates in Israel. Z. Schmiovitch et al. developed a semi-automatic system for maturity determination of fresh dates using NIR in which all testing was nondestructive. This has been trialed in commercial packinghouses as a method of acceptance or rejection of a load of dates.

  • Green tomato fruit can range in maturity from immature-green to advance mature-green. The concept tested by researchers at the University of Georgia (Chi N. Thai et al.) is to combine NIR to measure chlorophyll content and couple this information with x-rays to then determine the maturity of green tomatoes. This work is underway.

My Observations about NIR
NIR is a developing technology whose full potential has yet to be realized.

  1. Japanese peach, nectarine, and apple packers are using NIR to nondestructively determine soluble solids and, to some extent, skin chlorophyll content of fruits. This practice will probably increase and most likely will influence grower returns.

  2. Currently commercially available technology using NIR reflected light can determine soluble solids and skin chlorophyll.

  3. In the future, the technology is likely to improve to include the measurement of transmitted light, which would allow the measurement of acidity, watercore, and eventually internal defect sorting. Nondestructive sorting for firmness using NIR alone is not a near-future option.

  4. Japanese consumers require fruits which are firm and totally free of blemish and decay. Provided these criteria are met, soluble solids sorting may help Washington packers provide Japanese consumers with the type of apples they currently demand.

  5. As a result of this trip, I advise the Washington apple industry to:

    • Employ an experienced and qualified NIR agricultural engineer to work with a horticulturist to evaluate the possible uses and feasibility of NIR. This is a very difficult area to work on, as the engineering aspects are complex.

    • Determine the availability of Japanese NIR sorting equipment to USA commercial interests. In addition, determine whether this type of equipment is being developed in the USA. Communication with US manufacturers of NIR equipment and packingline suppliers as well as agricultural engineers familiar with NIR technology would help delineate options. (This trip did not guarantee the availability of NIR sorting equipment.)

    • If appropriate, obtain a laboratory bench to permit testing of this technology on Washington fruit.

      The question of whether NIR data can help Washington packers determine fruit "quality" must be answered. SS measurement alone will not provide valuable information on the firmness of Red Delicious. Although firm fruit have high SS levels, less firm (esp. overripe) fruit can also have high levels of SS.

      Skin chlorophyll levels might provide an indication of maturity (i.e., ground color), but this does not directly relate to firmness.

    • Providing laboratory tests prove positive, a packinghouse should be recruited to use this equipment on trial lots of fruit. Recognize the current throughput is 3 fruits per second per line and there will be a reduction in accuracy at faster speeds. No information was available on how much of a reduction in accuracy would be experienced at the speed used in most packinglines.

    • Communication with NIR technologists (industry and scientists) should be opened so the Washington industry can remain apprised of progress. For example, I am aware that a New Zealand group is working on NIR and has sent a scientist to work on this technology in a laboratory in Japan for a full year.

    • This should be considered a long-term research project as currently there are strong limitations to throughput (fruit orientation and accuracy of measurement) and meaningfulness of the data once it is collected. It is obviously not equipment which can be purchased off the shelf.

Selected References of NIR Research

Annotated bibliography on the use of light in assessing the quality, maturity or defects in fruit and vegetables 1967-1974. 1975. Commonw. Bur. Hortic. Plant Crops. 15/75. 2 p.

Birth, G.S., G.G. Dull., W.T. Renfroe and S.J. Kays. 1985. Nondestructive spectrophotometric determination of dry matter in onions. JASHS 110 (2): 297-303.

Bochereau, L., P. Bourdgine and B. Palagos. 1992. A method for prediction by combining data analysis and neural networks; application to prediction of apple quality using near-infrared spectra. J. Agric. Eng. Res. 51(3): 207-216.

Dull, G.G. 1986. Nondestructive evaluation of quality of stored fruit and vegetables. Food Tech. 40 (5): 106-110.

Dull, G.G. 1991. Nondestructive determination of moisture and sugars in dry and semi-dry fruits by near infrared spectrophotometry. BARD. 31 pp.

Dull, G.G., G.S. Birth and R.G. Leffler. 1989. Use of near infrared analysis for the nondestructive measurement of dry matter of potatoes. Amer. Pot. J. 66: 215-225.

Dull, G.G., G.S. Birth, D.A. Smittle and R.G. Leffler. 1989. Near infrared analysis of soluble solids in intact cantaloupe. J. Food Sci. 54 (2): 393-395.

Dull, G.G., R.G. Leffler and G.S. Birth. 1995. Nondestructive measurement of soluble solids in fruits having a rind or skin. USDA Patent. 1995. 1 p.

Dull, G.G., R.G. Leffler, G.S. Birth and D.A. Smittle. 1990. Instrument for nondestructive measurement of soluble solids in honeydew melons. Pap. Amer. Soc. Agric. Eng. Winter 1990 (90-3553) 6 p.

Kawano, S. 1994. Present condition of nondestructive quality evaluation of fruits and vegetables in Japan. JARQ 28: 212-216.

Kawano, S., T. Fujiwara and M. Iwamoto. 1993. Nondestructive determination of sugar content in Satsum mandarin using near infrared transmittance. J. Japan Soc. Hort. Sci. 62 (2): 465-470.

Kawano, S., H. Watanabe and M. Iwamoto. 1992. Determination of sugar content in intact peaches by near infrared spectroscopy with fibre optics in interactance mode. J. Japan Soc. Hort Sci. 61 (2): 445- 451.

Korcak, R.F., V. Walker and K.H. Norris. 1990. Measurement of fruit tree total leaf nitrogen by near infrared reflectance spectroscopy. Acta Hort. ISHS. 274: 241-247.

Singh, N. and M.J. Delwiche. 1994. Machine vision methods for defect sorting stone fruit. Trans ASAE 37(6): 1989-1997.

Throop, J.A., D.J. Aneshansley and B.L. Upchurch. 1992. Near IR and color imaging for bruise detection on Golden Delicious apples. Final report on a cooperative agreement between USDA-ARS and Cornell University 1992. No. 58-1931-1-125.

Upchurch, B.L. 1993. Photometric imaging for detecting surface and internal defects of apples. BARD. 68 pp.

Upchurch, B.L., J.A. Throop and D.J. Aneshansley. 1994. Influence of time, bruise-type and severity on near-infrared reflectance from apple surfaces for automatic bruise detection. Trans. ASAE 37(5): 1571-1575.

Upchurch, B.L., J.A. Throop and D.L. Aneshansley. 1993. Influence of time and bruise type on near infrared reflectance for automatic bruise detection. Pap Am. Soc. Agric. Eng. 93-3596. 14 p.

Dr. Eugene Kupferman, Postharvest Specialist

WSU Tree Fruit Research and Extension Center
1100 N. Western Ave., Wenatchee, WA 98801
Kupfer@wsu.edu

Tree Fruit Postharvest Journal 8(2):4-9
June 1997

Tree Fruit Research & Extension Center, 1100 N Western Ave, Washington State University, Wenatchee WA 98801, 509-663-8181, Contact Us