Cleaning, cleanliness, and hygiene are all undeniably central factors in human well-being, impacting both sensory experiences and the mitigation of infection spread. With the ongoing societal shift towards increased digitalization and automation, there is a growing interest in leveraging technology to rationalize and improve cleaning processes. Cleaning management systems utilizing connected sensors have emerged to optimize cleaning schedules in diverse settings such as restaurants, healthcare facilities, offices, shops, warehouses, and public spaces. This literature review report explores the vast landscape of publications addressing the topic of sensors that may provide viable data for decision-making algorithms in cleaning management systems. It also includes a few illustrating examples of commercially available sensors, as well as a list of some cleaning management systems on the market that employ sensors and data. Such systems may for instance exploit consumption data, detect visitor presence, and are often utilizing gas sensors. Gas sensors, being able to measure air quality, are maybe of particular interest for improving cleaning of restrooms. Some may be able to detect pure gases (like carbon dioxide), some react to several gases, and some are able to capture complex odours. Literature indicates that advancements in nano materials will likely lead to stable, energy-efficient gas sensors with increased sensitivity and selectivity. Electronic noses, arrays of gas sensors trained by algorithms, enable detection of complex odours and may enhance the selectivity of gas sensors. However, challenges persist, urging the need for continued improvement in gas sensor technology and smell sensing solutions, especially if to be integrated into Internet of Things (IoT) solutions. The review outlines various principles underlying gas sensors, emphasizing the preference for stable, low-energy technologies in IoT applications. While chemiresistive sensors exhibit promise, historical limitations in their sensitivity and selectivity have led to preferring other principles for certain applications, such as non-dispersive infrared and photoacoustic spectroscopy sensors. The review also describes technologies available for detecting other phenomena, including optical, ultrasound, and radar sensors. It is also underscored the potential in combining multiple sensor technologies to comprehensively assess soiling, tailored to specific environmental contexts. Despite significant progress, literature indicates that much work remains to fully exploit the capabilities of these integrated techniques for timely and efficient cleaning in diverse settings.

A new concept has been developed for characterizing the real-time evolution of the three-dimensional pore and lamella microstructure of bread during baking using synchrotron X-ray microtomography (SRµCT). A commercial, combined microwave-convective oven was modified and installed at the TOMCAT synchrotron tomography beamline at the Swiss Light Source (SLS), to capture the 3D dough-to-bread structural development in-situ at the micrometer scale with an acquisition time of 400 ms. This allowed characterization and quantitative comparison of three baking technologies: (1) convective heating, (2) microwave heating, and (3) a combination of convective and microwave heating. A workflow for automatic batchwise image processing and analysis of 3D bread structures (1530 analyzed volumes in total) was established for porosity, individual pore volume, elongation, coordination number and local wall thickness, which allowed for evaluation of the impact of baking technology on the bread structure evolution. The results showed that the porosity, mean pore volume and mean coordination number increase with time and that the mean local cell wall thickness decreases with time. Small and more isolated pores are connecting with larger and already more connected pores as function of time. Clear dependencies are established during the whole baking process between the mean pore volume and porosity, and between the mean local wall thickness and the mean coordination number. This technique opens new opportunities for understanding the mechanisms governing the structural changes during baking and discern the parameters controlling the final bread quality. © 2023 The Author(s)

This report reviews various measurement techniques and methods for assessing meat quality in lamb and beef after slaughter. It has mainly been based upon searches in databases of scientific literature, but also on discussions with industry people and colleagues.

The concept of meat quality is multifaceted and not easily defined. At the same time, most people who appreciate a good piece of meat would agree on that tough meat should be avoided. Hence, one property that has attained considerable attention in the literature is tenderness. Another characteristic that is usually emphasized as something positive is marbling, partly due to several studies that have indicated a relation between marbling and taste, juiciness and tenderness, all being properties that are important for a positive eating experience. Other features that attract attention are e.g. fatty acid composition, water retention ability, pH and colour.

The first attempts to find an objective method for tenderness measurement were based on mechanical methods. Essentially a spear-like object that was shot into the meat, while the resulting force was measured. For measurement of other parameters, such as pH, relatively traditional techniques were similarly used, based on more or less analogue technology. However, in recent decades there has been an increasing amount of studies using techniques that have benefited from the exponential development of digital and solid-state technologies. This development has e.g. led to easier ways to generate, measure and analyse electromagnetic, optical and acoustic signals.

A common approach in simpler measurement methods is trying to find an algorithm that is based on analysis of the frequency response of a signal that typically may be of electrical nature, ultrasound or light. Methods utilizing NIR have been particularly promising. One example of NIR equipment is NitFom^TM, which is used for quality assessment of fat in pork meat. Methods based on measurement of electrical impedance have also, at least periodically, found establishment on the market.

There are also more advanced approaches, with the ambition to obtain spatial resolution of properties within the object under assessment. For measurements with three-dimensional resolution, primarily computer tomography imaging (CTI) and magnetic resonance imaging (MRI) are candidates, but also ultrasound. Hitachi-Aloka, makes ultrasonic equipment that can be used for scanning live animals, and another example of ultrasonic equipment is ECM EXAGO. MRI and CTI, however, are still too expensive, advanced and slow to be realistic alternatives for online measurement in the industry in the near future.

MRI and CTI are developments of NMR (Nuclear Magnetic Resonance) and X-ray. While also NMR seems to be a little too expensive for industrial applications, and primarily a lab tool, X-rays have been used in the meat industry since the 1970s, e.g. for measurement of fat content. One commercial product that early found establishment on the market is Anyl-Ray Oystar.

The development of the digital camera opened up for advanced image processing. There are several studies based on analysis of the kind of information that can be extracted from RGB images, so-called vision technology, but also on analysis of images that contain much more detailed spectral information, so-called hyper- or multispectral image analysis. However, analysis of the amount of information that is collected with the latter kind of technology requires large computation and data management capabilities. In this context, the continuing development towards more accessible computational power is highly valued.

While image processing based on RGB information has been successful in measuring more or less what is also perceived by human eye (in this context such parameters as marbling and colour), hyperspectral image analysis has shown potential to go one step further. In addition to visual properties, the technology has shown promising results in measuring such things as chemical composition (e.g. proportions of fat, protein and water), pH and tenderness. Much resources have been invested in development of functional systems for online classification of meat in the industry. The results have been promising, and companies have been started up for the purpose, but the definitive breakthrough has not yet taken place.

In conclusion, several attempts have been made to find objective measurement methods for assessing and potentially classifying meat quality. Many promising results have been reported in the literature. Yet it is difficult to make any recommendations on one single salvaging technique based upon these results. Possibly, the technology that is currently attracting the most attention and hopes is hyperspectral image analysis, especially if the intention is to find a technology, suitable for forming the basis for a classification system. In such a context, hyperspectral imaging is a technology that meets many positive criteria: it is contact-free, it has spatial resolution, it combines advantages of both vision and NIR. There are also several studies that show promising results, and there is still good hope that the technology will develop further in near-time (both in terms of price and performance), hand in hand with the trend in society towards increased digitalisation (i.e. development of artificial intelligence, better and cheaper sensors, increased access to computational power, connected devices, etc.).

I arbetet att ta fram denna sammanställning av olika mättekniker och metoder för att bedöma kvalitet på lamm- och nötkött efter slakt, har vi främst baserat oss på sökningar i databaser över vetenskaplig litteratur. Utöver detta är den ett resultat av diskussioner med branschfolk och kollegor.

Begreppet köttkvalitet är mångfasetterat och inte så lätt att definiera. Samtidigt känner de flesta som uppskattar en god köttbit till att man gärna vill undvika ett segt kött, så en egenskap som ofta har fokuserats på i litteraturen är mörhet. En annan egenskap som brukar framhävas som positiv är marmorering, bland annat eftersom det finns studier som visar att köttets marmorering har ett samband med både smak, saftighet och mörhet, vilka är viktiga egenskaper för en positiv ätupplevelse. Fler egenskaper som rönt uppmärksamhet är bland annat fettsyresammansättning, vattenhållande förmåga, pH, och färg.

De första försöken att finna en objektiv metod för att mäta mörhet byggde på mekaniska mätmetoder, i stort sett ett spjutliknande föremål som sköts in i köttet, samtidigt som en kraft uppmättes. Mätning av andra parametrar, så som pH, byggde på liknande sätt på mer eller mindre traditionell analog teknik. De senaste decennierna har det dock tagit fart med studier på mätmetoder som gynnats av den exponentiella utvecklingen av digitala tekniker och halvledarteknik. Den utvecklingen har bland annat lett till enklare sätt att generera, mäta och analysera elektromagnetiska, optiska och akustiska signaler.

Vanligt angreppssätt i enklare mätmetoder är att försöka finna en algoritm som bygger på analys av frekvensresponsen vid mätningar av impedans, ultraljud eller ljus. Särskilt lovande bland den typen av mätmetodik har NIR varit. Exempel på NIR-utrustning är NitFom^TM som används för kvalitetsbedömning av fett hos gris efter slakt. Även metoder baserade på mätning av elektrisk impedans har åtminstone periodvis funnit etablering på marknaden.

I andra änden finns mer avancerade angreppssätt, med försök att erhålla rumslig upplösning av mätobjektet. Vid mätning i 3D är det främst datortomografi och MRI (Magnetic Resonance Imaging – magnetisk resonansavbildning) som är aktuella, men även ultraljud. Som exempel på tillverkare av ultraljudsutrustning kan nämnas Hitachi-Aloka, som bland annat säljer utrustning för skanning av levande djur. Ett annat utrustningsexempel är ECM EXAGO. MRI och datortomografi framstår dock ännu som för dyra, avancerade och långsamma för att i närtid vara realistiska alternativ för online-mätning inom industrin.

MRI och datortomografi är egentligen utvecklingar av NMR (Nuclear Magnetic Resonance – kärnmagnetisk resonans) och röntgen. Medan även NMR fortfarande verkar vara lite för kostsam för industriella tillämpningar, och främst ett labb-verktyg, så har röntgen däremot ända sedan 70-talet använts inom slaktindustrin, bland annat för att mäta fettinnehåll – exempel på kommersiell produkt är Anyl-Ray Oystar.

Utvecklingen av digitalkameran öppnade för avancerad bildbehandling. Det finns flera studier baserade på den typ av information som finns i vanliga RGB-bilder, s.k. vision-teknik, men även på information i form av avbildningar med mycket mer detaljerat spektralt innehåll, s.k. multi- eller hyperspektral bildanalys. Att analysera all information som samlas in med den senare typen av teknik kräver dock stora beräknings- och datahanteringsresurser, så där är utvecklingen mot alltmer lättillgänglig dator- och beräkningskraft välkommen.

Medan bildbehandling baserad på RGB-information har haft framgång att mäta sådant som uppfattas av mänskliga ögat, så som marmorering och färg, så har hyperspektral bildanalys visat på potential att nå ett steg längre. Utöver visuella egenskaper har tekniken givit lovande resultat att mäta sådant som kemisk sammansättning (andel fett, protein, vatten etc.), pH och mörhet. Mycket resurser har satsats på att ta fram fungerande system för klassificering av kött inom industrin, resultat har varit lovande, och företag startats för ändamålet, men ännu har inte det definitiva genombrottet skett.

Som framgår ovan så har det gjorts åtskilliga försök att finna objektiva mätmetoder för att bedöma, och potentiellt klassificera, köttkvalitet. Många lovande resultat har rapporterats i litteraturen, ändå är det svårt att enbart utgående från den landa i klara rekommendationer om en enda saliggörande teknik. Kanske är den teknik som just nu tilldrar sig mest uppmärksamhet och förhoppningar hyperspektral bildanalys, och särskilt om man tänker sig att tekniken ska utgöra grund för ett klassificeringssystem. Det är en teknik som uppfyller många positiva kriterier: den är beröringsfri, den har rumslig upplösning, den kombinerar fördelar hos både vision och NIR, det finns flera studier som visar på lovande resultat, och det finns fortfarande gott hopp om att tekniken kommer utvecklas i närtid både pris- och prestandamässigt hand i hand med digitaliseringstrenden (d.v.s. utvecklingen av artificiell intelligens, bättre och billigare sensorer, ökad tillgång till datorkraft etc.) i samhället.  

FREEZEWAVE project is an ERANET SUSFOOD project (May 2015 - Dec. 2018). It concerns the freezing of foods using a highly innovative technique combining freezing at slow rate with part time emission of microwaves (2450 MHz); this innovative concept has been investigated recently by ONIRIS and results showed that a 62% decrease of the average ice crystal size was acquired when samples were frozen under microwave irradiation compared to the control (Xanthakis et al., 2014, IFSET - study on pork meat). To improve the quality of frozen foods, fast freezing is usually recommended, resulting in a reduction of the ice crystals. However, a fast freezing increases the energy demand: low ambient temperature & high air velocity to enhance the rate of heat transfer are needed to achieve a fast freezing. FREEZEWAVE project aims at expanding & optimizing the concept to several foods (sauce, meat, vegetable & ready to eat meals) and also at designing industrial equipment. The novel concept concerns the freezing equipment sector thanks to a French SME partner (SAIREM) and the global frozen food sector. FREEZEWAVE will provide scientific knowledge and new scientific insights in food freezing. Project's outcomes may also be of interest for non-food applications such as biotechnology.

A pilot-scale process for continuous in-flow microwave processing of particulate pumpable foods, designed and implemented at RISE Agrifood and Bioscience, was evaluated for heat treatment of a particulate, viscous model food at high-temperature conditions at 2450 MHz. The microwave system has three consecutive cavities, one excited by the TM020 microwave mode that heats primarily in the centre of the tube, and two cavities fed by TM120 modes that heat primarily in the tube periphery. In this paper, combined TM020 and TM120 tubular microwave heating is evaluated as an alternative to high-temperature short-time (HTST) processing for a high-concentrated particulate model product. Rapidness in heating of the product was evaluated after tubular microwave heating for different time-temperature conditions, corresponding to the required microbiological inactivation for a model product intended for storage at ambient conditions. Moreover, the effects on product quality of the microwave heated model soup were investigated in terms of texture, piece integrity and colour. Microstructural analysis was used to gain an understanding of the effects of heating at a microscopic scale. It was found that the microwave-assisted HTST system results in large process flexibility. Additionally, it offers advantages in product quality in terms of piece integrity and texture.

Dielectric properties of a variety of foodstuffs were measured during freezing and thawing conditions, and are presented here as functions of temperature. Measurements were done with the Agilent 85070 high temperature dielectric probe. The differences in characteristics of the graphs resulting from cooling and thawing are discussed. A specific model accommodating the amount of frozen water in the food sample as a function of temperature is used and compared to experimental results. The mass fraction of each component is coupled to the dielectric properties of each fraction to evaluate the effective dielectric properties based on a two phases Maxwell-Wagner model considering ice as a dispersed fraction and food as a continuous fraction. The unknown dielectric properties of dry matter and an adjusting coefficient are used to optimize the fitting between experimental data and calculated data.

A pilot-scale process for continuous in-flow microwave processing of foods, designed and implemented at SP Food and Bioscience, was evaluated for heat treatment of a homogeneous model food for high-temperature short-time (HTST) conditions, at constant total input microwave power, at 2450 MHz. The microwave system has three consecutive cavities, one excited by the TM₀₂₀ microwave mode that heats primarily in the tube centre, and two TM₁₂₀ mode cavities that heat primarily in the tube periphery. The temperature uniformity of the homogeneous model food after microwave heating is here evaluated in terms of spatial distribution, for different set-ups of input microwave power in each cavity and for different order of the placement of the cavities, while maintaining the total input microwave power. The microwave heating uniformity is evaluated, based on measured and calculated radial temperature profiles. Combined TM₀₂₀ and TM₁₂₀ heating was found to result in more uniform heating by means of spatial temperature uniformity over the tube cross section. Furthermore, appropriately selected microwave power distribution between the centre and periphery heating cavities results in a stable heating profile in the studied food, that differs only about 10 °C or less between highest and lowest average values directly after microwave heating.

There is currently a need in developing new decontamination technologies for spices due to limitations of existing technologies, mainly regarding their effects on spices’ sensory quality. In the search of new decontamination solutions, it is of interest to compare different technologies, to provide the industry with knowledge for taking decisions concerning appropriate decontamination technologies for spices. The present study compares infrared (IR) and microwave decontamination of naturally contaminated paprika powder after adjustment of water activity to 0.88. IR respectively microwave heating was applied to quickly heat up paprika powder to 98°C, after which the paprika sample was transferred to a conventional oven set at 98°C to keep the temperature constant during a holding time up to 20 min. In the present experimental set-up microwave treatment at 98°C for 20 min resulted in a reduction of 4.8 log units of the total number of mesophilic bacteria, while the IR treatment showed a 1 log unit lower reduction for the corresponding temperature and treatment time. Microwave and IR heating created different temperature profiles and moisture distribution within the paprika sample during the heating up part of the process, which is likely to have influenced the decontamination efficiency. The results of this study are used to discuss the difficulties in comparing two thermal technologies on equal conditions due to differences in their heating mechanisms

During infrared processing, transmitted energy is attenuated exponentially with penetration distance, and its intensity is gradually lost while passing through absorbing or scattering media. Penetration depth is a complex function of chemical composition of a food product, its physico-chemical state and physical properties and wavelength spectrum of energy source. Knowing penetration depth leads to better designing commercial sterilization processes for food products like spices. Therefore, the objective of this study was to determine penetration depth of various spices (paprika powder, black pepper and oregano) as a function of water activity under infrared processing conditions. For this purpose, spice samples were prepared at various water activity (aw) levels, and heat flux measurements were carried out to determine the penetration depth. Penetration depth was determined to increase with increasing aw for black pepper seeds and paprika powder while there was no significant change for oregano leaves as a function of aw. Knowing penetration depth is important to design an effective infra-red processing system and an important issue for surface pasteurization processes since infrared radiation effects on microbial inactivation decrease with sample thickness