International Scientific Conference era-6

Скачать 374.53 Kb.
НазваниеInternational Scientific Conference era-6
Дата конвертации26.10.2012
Размер374.53 Kb.
  1   2   3   4   5   6   7   8   9   ...   20

powerpluswatermarkobject14981027International Scientific Conference eRA-6

Control & Automation Session

1. Non-Destructive Fish Quality Control: A Novel Stochastic Model-based Scheme for Cost-Effective Freshness Evaluation

D. Dimogianopoulos1, K. Grigorakis2

1 Dpt. of Automation, Technological Educational Institute (T.E.I) of Piraeus, Athens, Greece
Tel: +30 2105381183, E-mail:

2 Institute of Aquaculture, Hellenic Centre for Marine Research, Agios Kosmas, Athens, Greece

Tel: +30 2109856723, E-mail:


A novel stochastic model-based scheme for reliably evaluating fish freshness well before spoilage is significant is introduced. Fish samples are tested via a vibration-like procedure, and their resulting response is accurately captured (under typical test-related uncertainties) via stochastic models. These provide specific indicators characteristic of the sample’s texture, itself closely related to its freshness. Thus, the unknown freshness of a tested fish may be evaluated by comparing its indicators with the nominal “fresh-fish” values. This comparison relies on a custom-built statistical hypothesis test, capable of issuing reliable decisions while taking into account the risk of the decision-making process. Tests involving 0-, 3- and 6-day-old fish samples show very promising results.


Freshness is a major indicator of fish quality, with its loss characterizing the end of high quality shelf life of fish, that is, reduced market price. Hence, early freshness evaluation is critical for the respective aquaculture and fish industry. Traditionally, freshness is assessed via organoleptic (or sensory) methods describing how freshness is perceived by the human senses, and/or physicochemical evaluations of the changes occurring in fish at physical and chemical level. Although effective and reliable, these methods impose destruction of the test sample, and require significant time, tooling and expertise from the test-conducting personnel. As such, they may prove expensive and not ideally suited to day-to-day industrial practice.

Such reasons have led to the development of various rapid non-destructive techniques and tools for evaluating freshness. These include sensors for rapid measurement of fish (or meat) dielectric properties which change over storage (potentiometric techniques) [1]-[5], electronic gas sensors measuring changes of volatile fish products altered with freshness reduction [6], [7], and so on. Although such tools are well suited to industrial use due to their rapidity and non-destructive implementation, they do exhibit weaknesses related to their considerable cost and complexity. However, their key limitation is the low sensitivity in detecting early post mortem changes [8]. For instance, the tool in [3] distinguishes between “spoiled” and “unspoiled” meat products. For fish products, this would translate into high effectiveness in detecting the spoilage degree (that is, how close a fish is for being discarded as unacceptable), but low ability in distinguishing between fish of various early freshness stages. This results from the dielectric properties and the volatile compounds of fish changing radically only when bacterial spoilage has already advanced ([9], p. 132). Hence, apart from being time- and cost-effective, a scheme evaluating fish freshness should be sensitive to early post-mortem changes.

This work introduces a cost-effective evaluation scheme for onsite industrial use, which does not resort to (destructive) organoleptic tests, complicated electronics or special sensors for reliably evaluating early post-mortem fish freshness. Relevant fault diagnosis methodologies (as applied to various cases from aircraft systems [10] to medical orthopedic implants [11]) are used for defining the scheme’s concept of operation. The scheme operates on a system formed by the fish sample under inspection, a spring balance, a small plastic hammer and an electro-pneumatic transducer transforming the balance readings into voltage. Placed on the balance pan, the fish sample suffers a hit from the hammer (dropped from a specific height) and the ensemble vibrates to a stop. The vibrating system’s resulting force signal –measured and sampled by a digital oscilloscope- is modeled via discrete-time output-only stochastic AutoRegressive (AR) representations. These accurately describe the system dynamics under slightly offset hits, noise in measurement and other uncertainties. The damping factors and natural frequencies of the identified AR models are key indicators of the sample’s texture, itself directly linked to its freshness. Thus, testing a system involving an “unknown-freshness” fish sample, and statistically comparing its key indicators with those from a system with a “fresh-fish” sample allows for evaluating the -hitherto unknown- freshness.

The current scheme presents several improvements upon previous relevant work [8]: The hit (excitation) is, now, allowed to be far stronger than the low-height step-like hit in [8], resulting to better noise-to-signal ratio of the response signal. The identified AR models may be decomposed into parts accounting for major signal dynamics and into parts representing (high frequency) noise. Hence, one may only monitor that part of the dynamics affected by the changed freshness. Finally, the use of statistical hypothesis tests for issuing decisions on the tested sample’s freshness means that, now, the risk of “wrong decision” during the decision-making process is quantified. The paper is organized as follows: Section 2 presents the test tooling, the stochastic AR model identification, and the design of the hypothesis test in detail. Section 3 shows tests with six fish samples (0-, 3- and 6-day-old fish). The test results and relevant discussion, along with a short cross-validation by means of an organoleptic evaluation of freshness of the tested fish samples are presented in section 4. Finally, some concluding remarks are given in section 5.

  1   2   3   4   5   6   7   8   9   ...   20

Добавить в свой блог или на сайт


International Scientific Conference era-6 iconInternational Scientific Conference Committee

International Scientific Conference era-6 iconXix international Scientific and Engineering Conference on Photoelectronics and Night Vision Devices

International Scientific Conference era-6 iconVi international scientific and practical conference for memory of P. Roudik
Международная научно-практическая конференция по психологии спорта и физической культуры памяти П. А. Рудика «Рудиковские чтения...

International Scientific Conference era-6 iconХалықаралық ғылыми-тәжірибелік конференция материалдары 9-10 маусым 2011ж
Репродуктивное здоровье и гинекология. Проблемы и перспективы: Материалы международной научно-практической конференции. Reproductive...

International Scientific Conference era-6 iconIx ukrainian Scientific and Technical Conference

International Scientific Conference era-6 iconInternational Conference on Malignancies in aids and Other Immunodeficiencies April 28-29,2003 Natcher Conference Center Bethesda, Maryland
Ш international Conference on Malignancies in aids and Other Immunodeficiencies April 28-29,2003 Natcher Conference Center Bethesda,...

International Scientific Conference era-6 iconInternational Scientific Advisory Committee International Organizing Committee

International Scientific Conference era-6 iconInternational Scientific Committee

International Scientific Conference era-6 icon729 22. 10 2048 era athens 1 all era stations heard in parallel kong4

International Scientific Conference era-6 iconT he Nation in the Global Era Conference Abstracts Pace University, New York, ny friday, June 6: 12: 45 – 2: 30: Friday Panels Session a transnational Labor/Capitalist Relations

Разместите кнопку на своём сайте:

База данных защищена авторским правом © 2012
обратиться к администрации
Главная страница