NL1041809A

NL1041809A - A spectral imaging system to detect contamination.

Info

Publication number: NL1041809A
Application number: NL1041809A
Authority: NL
Inventors: Beijersbergen Marco; Brusasca Marta; Krijger Joris; Rolloos Jan
Original assignee: Condi Food B V
Priority date: 2015-10-09
Filing date: 2016-04-11
Publication date: 2017-04-24

Abstract

Abstract: The current invention consist of an optical inspection categorizes system that and quantifies detects non-destructively, contaminations (defined as presence of unwanted substances) in real time on biological and non-biological surfaces. The most important application is the detection of bacteria by using auto-fluorescence. By using a reference database of optical footprints for a variety of contaminants the current innovation can enable rapid bacteria detection.

Description

Applicant: Condi Food B.V. te Warmond, Nederland Inventors: Marco Beijersbergen te Sassenheim, Marta Brusasca te Leiden, Joris Krijger te Leiden, Jan Rolloos te Noordwijk Title: A spectral imaging system to detect contamination

Description

Summary of the Invention

The present innovation develops the existing techniques to such specificity that an operational device for this purpose is produced. An example, but not limited to it, is the forensic industry where biological traces as blood DNA and hair samples are analysed only qualitatively: the presence or absence of a particular substance is only indicated without showing the exact chemical composition. With the current innovation an in situ chemical analysis of the object to be measured is possible. Another example is the food industry where, in order to detect the presence of unwanted bacteria, a laboratory analysis is necessary and the food needs to be stored multiple hours before the results are communicated.

The current invention presents a solution which can reduce that stand-by time drastically by immediately giving the spectral fingerprint of contaminants from which, via algorithms, the microbiological contamination of the measured object can be deduced. The innovation offers an advanced sensor platform that combines hyperspectral imaging with embedded computer and customized scientific software, which makes possible to measure hygiene, via optical inspection methods, in order to prevent contamination related diseases and infections. Proactive direct observation of food condition can enable the use of monitoring techniques to identify and predict safety issues .

Field & Background

In the broadest sense this invention is related to the field of optical inspection systems applicable in multiple domains. More specifically this invention is considered a method and apparatus to detect and quantify microbiological contamination. Although the laboratory testing, which is the most common technique for bacteria detection, remains highly reliable it is also a costly technique in time and labour. In the food industry, time spent in waiting for the bacteria analysis means a costly stalling of products. In healthcare, where it can take days to elaborate the results, this delay can lead to unnecessary costs for a temporary solution, such as broad spectrum antibiotics instead of narrow spectrum ones. The current innovation wants to overcome these drawbacks by using an optical measurement system to detect microorganisms that is continuous, non-destructive and in real time. By using spectroscopic measurements the hygiene, in this case defined as the absence of unintended matter in an object to be measure, can be determined. An example, but not limited to this, could be the presence of cleaning product residual in recycled product packages.

The detection of physical and chemical properties can be used for a wide array of applications. Although the usage of spectral imaging methods for different samples is not unique, the combination of camera developed for space applications combined together with the concentrated computer power open up the unique possibility for continuous real-time, optical, non-destructive inspection. In the current state of the art there is no device that can give direct real time feedback on the presence or absence of (unwanted) microbiological organisms.

Advantages

The current invention tackles the problems of current microorganism detection methods. No system so far can optically monitor an entire object or sample in vivo, without using reagents (e.g. dyes). This real time measuring of microorganisms means there is direct feedback and no waiting time to obtain the results. The nondestructive nature of the invention, resulting from the redundancy of reagents or markers, brings forward the possibility for continuous measuring as opposed to sample based measurements. This means that in the present invention no marker for the excitation of luminescence molecules has to be added to the object that is going to be measured. The hyperspectral images are acquired on reflection and a large area can be scanned at once. The current invention detects the auto fluorescent emission using highly sensitive optical sensors, intense excitation light sources and appropriated filtering. The fluorescence spectral image is then analysed to discriminate the spectral features that characterize a particular microorganism.

Therefore, because the composition of the whole surface of the object can be directly measured, the limitation of having sampling point measurement can be overcome. Ergo, the full product is scanned in real time without having to depend on limited point-like measuring points on the object to be monitored [as is for example the case with Pa. No. WO2014/180568 Al]. Small infected areas, as big as the pixel size of the camera, can be detected because of the 100% scanning of the surface of the product.

The application of this invention goes beyond the current monitoring of microorganisms and specifically bacteria in for example the medical sector as well as in the food industry. In these industries there are methods available for continuous in vivo monitoring of a potential bacterial infection site [Pub. No.:US2012/0143024 Al] but these methods can detect bacteria through fluorescence only if induced by markers .

The current innovation leaves behind the disadvantages of the previous methods by uniquely combining the continuous in vivo monitoring of the entire object that is to be measured with the non-destructive measurement; measuring without adding external materials as markers or dyes or any other intrusive exciting reagents.

Method

The innovation provides a method of inspection by detecting microbiological organisms. Simultaneously the following phases are part of the innovation: a) Acquire the (hyper) spectral image of the object to be measured. b) Analyse the spectral image to determine whether there is an unwanted substance present on the object/product. A comparison is done between the measured spectra and spectrum of the uncontaminated sample . c) In case a contamination is detected, the presence of typical molecules is identified. The fingerprint of the micro-organism is then evident and it can be classified. d) Determine if the micro-organism state is still living or already dead e) Determine what type of micro-organism is present, differentiating between prokaryotes (bacteria and archaea) and eukaryotes (for example fungi, parasites or plants ) . f) Within the detection of prokaryotes and specifically bacteria the present invention discriminate between domains. This means that every phylum will be distinguished and within it every class and subclass. g) In a final step distinction is made within a certain subclass. In order to identify the pathogen bacteria the micro-organism will be analysed to the level of bacteria serotype, the smallest possible level.

An example would be the detection of the dangerous salmonella Thompson bacteria on salmon. This caused a scandal in 2012 in the Netherlands. A scandal which can be prevented in the future with the current innovation. On a continuous basis the camera takes spectral images of all the salmons in the production line. In this particular application, the auto-fluorescence of certain bacterium molecules is used to discriminate among them.

Preferred Embodiment

Light

The inspection system works with a light source that illuminates the object to be measured and analyses the scattered light.

The light source is emitting either within the UV, the visible or IR range. The wavelength can be either monochromatic, indicating that only one wavelength is used, or pseudo monochromatic, meaning that more than one wavelength is used for the light source.

Camera Techniques

There are three types of spectral techniques that are suitable for this kind of measurements: • The hyperspectral camera: a camera that divides the spectrum in 20 up to 200 bands. • Multispectral camera: a camera which measures one or more specifically chosen broad bands. • Imaging spectrometer. A camera that measure more spectra with a higher resolution: 200 up to 2000 bands.

Housing

The system is preferably enclosed in light-tight housing to minimise the effect of straight lights from the surrounding environment. In this light-tight environment, calibration measurements need to be acquired once, or once in certain interval of time. However, in case it is not possible to guarantee a light-tight environment, calibration procedures need to be followed in order to establish the correct reference for the analysis.

Contamination Detection

The elaborated (hyper)spectral image shows on a 2D spatial image where the contamination is. The spectral data are used to discriminate between the different kinds of contaminants while the spatial coordinates inform about the location of the unwanted material. A second measurement (for example after a treatment) can reveal if the contamination is still present. The computational analysis of the results indicates which area is contaminated and for example which percentage of the surface is covered, and indicates the density or the concentration of the contaminants (g/cm2).

By first developing a model for the auto-fluorescence spectrum of the sample and reproducing the same measurement conditions, the deviation between the model and the measured sample is calculated. The deviation indicates whether contaminants are present which are not expected and/or wanted. A second model is used to analyses this micro-organism to determine which type and serotype it is comparing it to a reference micro-organism spectral database and to determine in which quantity and dispersion is present on the measured object.

So when a foreign microbiological presence is detected in the spectral image, an algorithm compares the datum to a reference database to determine the class, subclass and serotype of the measured contamination. Countermeasures can be taken immediately after the measurement to prevent safety issues.

The determination happens by means of measurements done on the spectral images of fluorescence, auto-fluorescence and, if present, contamination. This optical and contactless fluorescence comes to pass by shining light on a surface and analysing the reflecting light by means of a spectral camera. The reflected light of an organic material has a specific spectrum which can be compared to and differentiated from the spectrum radiated from the same material without contamination. This means that in the case of contamination two analysis are performed. The first computation, which is run on every object that is measured, analyses whether there is any deviation between the measured spectrum and the reference one (the same non contaminated material). In case a contamination is found, a second analysis is run to match the deviation with certain known contaminant spectra. The procedure for the measurement is to compare the measured object to the clean one and only when a significant deviation is detected a second comparison is run. This second comparison will look for a match in the contaminant database to determine what type of contamination, for example which serotype of bacteria, can match the measured spectrum in the investigated object.

Measured Objects

The detection of micro-organisms is possible on one or more biological and non-biological objects. With non-biological is meant material that does not live or has lived and that hasn't been taken from living material. The most prominent non-biological objects will be summed up beneath:

The detection of micro-organisms on:

Working surfaces, tools or machines.

Packing material.

Medical instruments Implants (such as silicones)

Floors, walls, ceilings and other forms of interior structures .

Surfaces for forensic applications (sheets, cars, clothes, fabrics etc.)

All non-biological agricultural material as for example artificial soil for mushrooms.

Tubes, pipelines, sewers, cranes, barrels or tanks. The detection of micro-organisms is also possible on biological surfaces .

With biological is meant everything that doesn't fall in the above category of non-biological. The most prominent biological objects will be summed up beneath:

Wooden constructions

Consumable goods: such as: meat, fish, poultry, sea food, vegetables, fruits and processed consumables such as ready to eat meals .

Resources such as barley, corn, fuel oil, glass, malt, rice, wheat, herbs and spices. Fluids: water, oil, diesel, vinegar and wine. Organic surfaces: humans and animals. For example on skin, wounds, or mucosa. Micro-organisms: the measuring of fungi and bacteria. This in order to determine whether there are pathogens present. For example unwanted fungi on cheese or souring bacteria in wine .

Criteria and Actions

These measurements on various surfaces open up the possibility to define criteria for hygiene. This can be binary criteria where a threshold is defined under which a product is considered clean and above the threshold it is considered contaminated. A more specific hierarchy of classification (for example a ranking from 1-10) can also be implemented. With this information a complete safety assessment can be done, assuring a 100% inspected product. This invention can result in immediate actions to confine any kind of contamination.

Figure Description and Example of Usage

Figure 1 depicts a non-limiting example of the setup used for the application showing the imaging system analysing a food sample.

Figure 2 is an example of the output of the hyperspectral camera. A monochromatic picture of a clean and contaminated sample can be reconstructed.

Referring to Figure 1, a spectral imaging system in accordance with an embodiment of the present invention is illustrated.

It consist of: 1) light source 2) Excitation filter 3) measured object 4) dicroic mirror 5) detection filter 6) hyperspectral camera 7) controller system 8) database

The operations of the invention runs as follows:

An excitation filter (2) selects ranges of wavelength of the light source (1) that illuminates the surface of the object to be measured (3), in this case (but not limited to) the surface of a fish. Detection means, in this example and hyperspectral camera (6) collect the reflected light from the product and through a detection filter (5) the light is filtered. The output data of the measurement systems is communicated to the controller system (7), which contains the referential data-bases (8). The comparison to reference tables and thereby the identification of targets, for example the salmonella bacteria, can be done automatically. A warning advise is given in case of positive detection.

Figure 2 shows how the differences between a contaminated product and a non-contaminated product become visible in the monochromatic image the optical measurement system records. In figure 2a a monochromatic picture at a particular wavelength of the measured object (1) is taken without any contamination while in figure 2b the measured object (1) has been contaminated. Detection of the contaminants (2) is possible with the device described in figure 1 and contamination is clearly visible on the measured object.

Claims

1) A spectral image registration system to measure pollution by using optical methods consisting of: a) A light source that emits light in the spectrum from UV to Near Infrared; b) Detection means for detecting the light reflecting from the measured object; c) At least one spectral filter element that selectively transmits a predetermined wavelength of light; d) A 'vision module' that focuses on the object to be measured and to record spatial and spectral data; e) Control means for scanning the object to be measured in one or more selected wavelengths; f) A reference database that collects the spectral properties of polluting substances; and, g) Analysis means to analyze found deviations and compare them with the reference database in order to determine what type of contamination is present on the measured object.

2) A spectral measurement system for measuring contamination according to claim 1 where the surface of the object to be measured is either organic or non-biological.

3) A spectral measurement system for measuring contamination as described in claim 2 where the measured reflected light is the auto-fluorescent emission of the measured object.

4) A spectral measurement system for measuring contamination according to claim 1 where the detection means consist of either a hyperspectral camera, or a multispectral camera or an imaging spectrometer.

5) A spectral measurement system for measuring contamination as described in claim 3 where the registration of the image and the analysis of the contamination are done in real time.

6) A spectral measurement system for measuring contamination as described in claim 5 where the entire area of the object to be measured is measured in one go.

7) A spectral measurement system for measuring contamination as described in claim 6 where the contamination consists of a bacterial contamination.

8) A spectral measurement system to measure contamination as described in claim 7 where the domain and subclass of the bacterium is determined to the level of the serotype.

9) A method for determining microbiological contamination using a system as described in claim 1, consisting of the following phases: a) Making a spectral image of the object to be measured; b) Analyzing the image to determine if there is abnormal or unwanted substance on the measured object; c) Detecting the presence of typical molecules that are characteristic of certain micro organisms; d) Determining whether these micro organisms are still alive or already dead; e) Determining the type of micro-organism present, in which a distinction is made between prokaryotes and eukaryotes; f) Within detecting prokaryotes and specifically bacteria, differentiating between the different domains to determine each 'phylum' and within it each class and subclass; and, g) Determining the serotype within the subclass to detect pathogenic bacteria.

10) A method as described in claim 9, which does not use markers for the optical detection of contamination.

11) The use of the method as described in claim 9, using a system as described in claim 1, in: a) agriculture and horticulture b) the food and commodities industry c) the pharmaceutical industry d) the medical field e) the construction f) the forensic field g) science h) transporters or transport i) any residual areas that do not fall under specification a through h.