Projects / Programmes source: ARIS

Complex hyperspectral system for automatic analysis and control of pharmaceutical pellet coating processes

Research activity

Code Science Field Subfield
2.06.00  Engineering sciences and technologies  Systems and cybernetics   

Code Science Field
T111  Technological sciences  Imaging, image processing 

Code Science Field
2.02  Engineering and Technology  Electrical engineering, Electronic engineering, Information engineering 
ANG NIR hyperspectral imaging, image analysis, automatic quality inspection, process analytical technology, contactless pellet analysis
Evaluation (rules)
source: COBISS
Researchers (21)
no. Code Name and surname Research area Role Period No. of publicationsNo. of publications
1.  18172  PhD Rok Bernard  Systems and cybernetics  Researcher  2011 - 2014  29 
2.  25528  PhD Miran Burmen  Systems and cybernetics  Researcher  2011 - 2014  107 
3.  21455  PhD Rok Dreu  Pharmacy  Researcher  2011 - 2014  290 
4.  28446  PhD Ilija German Ilić  Pharmacy  Researcher  2011 - 2014  139 
5.  33446  PhD Bulat Ibragimov  Systems and cybernetics  Researcher  2011 - 2014  43 
6.  29559  PhD Jaka Katrašnik  Systems and cybernetics  Junior researcher  2011 - 2012  16 
7.  21320  MSc Marko Knez  Computer science and informatics  Researcher  2011 - 2014 
8.  32850  PhD Matjaž Kosec  Systems and cybernetics  Researcher  2011 - 2014 
9.  02587  PhD Peter Kump  Physics  Researcher  2011 - 2014  240 
10.  15678  PhD Boštjan Likar  Systems and cybernetics  Head  2011 - 2014  381 
11.  27519  PhD Primož Markelj  Systems and cybernetics  Researcher  2011 - 2014  20 
12.  33166  PhD Uroš Mitrović  Computer science and informatics  Junior researcher  2011 - 2014  18 
13.  31279  PhD Miha Možina  Systems and cybernetics  Junior researcher  2011 - 2012  12 
14.  10677  PhD Marijan Nečemer  Chemistry  Researcher  2011 - 2014  348 
15.  06857  PhD Franjo Pernuš  Systems and cybernetics  Researcher  2011 - 2014  519 
16.  09031  PhD Stanko Srčič  Pharmacy  Researcher  2011 - 2014  674 
17.  29888  PhD Rok Šibanc  Pharmacy  Researcher  2011 - 2014  63 
18.  28465  PhD Žiga Špiclin  Systems and cybernetics  Researcher  2011 - 2014  135 
19.  20383  PhD Dejan Tomaževič  Manufacturing technologies and systems  Researcher  2011 - 2014  91 
20.  31986  PhD Peter Usenik  Systems and cybernetics  Junior researcher  2011 - 2013  15 
21.  23404  PhD Tomaž Vrtovec  Systems and cybernetics  Researcher  2011 - 2014  197 
Organisations (4)
no. Code Research organisation City Registration number No. of publicationsNo. of publications
1.  0106  Jožef Stefan Institute  Ljubljana  5051606000  88,443 
2.  0787  University of Ljubljana, Faculty of Pharmacy  Ljubljana  1626973  17,150 
3.  1538  University of Ljubljana, Faculty of Electrical Engineering  Ljubljana  1626965  27,564 
4.  2294  Sensum, sistemi z računalniškim vidom d.o.o. (Slovene)  Ljubljana  1554077  220 
Pharmaceutical pellets are made by agglomeration process in which fine particles are joined to improve content uniformity and flow characteristics, and to reduce segregation and dustiness. Additionally, pellets are typically coated with a thin film, serving as a protective layer against environmental factors, preventing the degradation of active pharmaceutical ingredients (API), controlling drug release rates, modifying color, masking taste, and defining the final size and shape of pellets. The highly important API release rates can be controlled by the volume-to-area ratios of pellets and with the compositions and thicknesses of the coating layers and/or composition of pellets matrix. Due to these advantages, coated pellets are increasingly replacing powder blends for filling pharmaceutical capsules and partially also in the process of pressing tablets. However, the highly complex multivariable pellet coating processes need to be controlled well to ensure high product yields and to prevent process failures, resulting in discarded batches and significant losses of revenues. For this purpose, a number of analytical methods are required for measuring the chemical (composition), physical (coating thickness) and geometrical (shape and size) properties of pellets. Currently, the average pellet size and coating thickness are commonly assessed off-line by microscopes or by dissolution tests, while the average chemical composition is usually verified by the HPLC (high performance liquid chromatography). Unfortunately, these methods are highly time consuming, retrospective, destructive, environment-harmful and provide no information on the spatial distributions of the pellet composition and film coating thickness, which both significantly affect the final therapeutical performance. Moreover, these classical methods are infeasible for real-time in-process pellet measurements, essential for effectively designing, controlling and optimizing the coating process and thereby assuring the high yield, quality and consistency of the final product – as stressed out in the PAT (Process Analytical Technology) initiative proposed by FDA (Food and Drug Administration) and EDA (European Drug Agency) as well. On the other hand, real-time statistical assessment of the average coating thickness can be achieved by the NIR (near infrared) point spectroscopy, while geometrical properties, such as size and shape, could be assessed by the analysis of NIR and visible hyper-spectral images, which can be acquired by the AOTF filters (acousto-optical tunable filters) or line-scan spectral cameras (see the attachment). However, in order to extract the geometrical, physical and chemical properties of pellets along with corresponding spatial distributions in real-time, new methods for automated analysis of spectra and hyper-spectral images need to be developed. In this manner we can achieve fast, efficient, reliable, noncontact, nondestructive, robust and environment friendly statistical analysis of pellets for reliable industrial process control.   In the proposed applied project, we will focus on the development of new methods for automatic, contactless and nondestructive real-time hyperspectral assessment of the geometrical, physical and chemical properties of pharmaceutical pellets. This will be performed by means of robust, automatic and real-time analysis of NIR and visible hyper-spectral images, especially focusing on assessing size, shape, composition, coating homogeneity and coating thickness of pellets. The development and integration of such methods will enable construction of innovative complex measurement systems and instruments, providing the means for efficiently designing, optimizing and controlling the production of pharmaceutical pellets and assessing the quality of pellets according to the PAT initiative, thereby assuring the quality and safety of medicines, increasing the productivity, and providing environment-friendly production.
Significance for science
The results of the research activities conducted during the applied project were disseminated in top-ranking scientific journals with high impact factors. Consequently, this will increase the visibility and reputation of Slovenia and in particular of the University of Ljubljana, as well as impact the long term development of technical and natural sciences in Slovenia. With respect to the nature of the applied project, we expect that the novel methods for in-line assessment of the most prominent pellet properties, such as the size and shape of pellets and the coating thickness and uniformity, will accelerate the transfer of top scientific knowledge to industry and stimulate novel research on the use of hyperspectral imaging technologies in the pharmaceutical production processes. As hyperspectral imaging is currently in the process of slowly but surely finding its way from the laboratory towards the pharmaceutical production processes, the conducted project also involved a fair share of basic research in the field of calibration and enhancement of acquired images and design of efficient illumination sources that are applicable to other fields of research, such as remote sensing, were hyperspectral imaging is considered an established imaging modality. Many of the findings will significantly affect future studies in the field of quality control by computer vision and enable substantial improvements of the existing imaging systems. In particular, extension of computer vision from the visible to the invisible part of the electromagnetic spectrum and the integration of multiple imaging modalities and spectral regions was shown to offer almost unlimited possibilities for the development of new technologies with high value added. Such systems will enable analysis, development and optimization of modern manufacturing processes and technologies in the pharmaceutical industry.
Significance for the country
The potentials of exploiting the results of the applied research project are numerous. The novel methods for real-time in-process assessment of the most prominent pelet properties will enable construction of modern high-tech control systems and instruments for simultaneous monitoring, understanding and optimization of pharmaceutical pellet coating processes, which will result in higher quality and safety of products, higher productivity and environment-friendly production. The prototype imaging system developed during the project presents a huge step towards efficient real-time in-process monitoring of the pharmaceutical coating processes. The results of the thorough evaluation based on multiple coating processes showed that the prototype imaging system and the novel image processing methods developed during the project, present a solid foundation for development of state-of-the-art systems suitable for application in full industrial scale coating processes. Besides, we expect that the novel methods for simultaneous calibration and resolution enhancement of hyperspectral images will promote addition research in the field. The image calibration and enhancement methods are of the utmost importance in industrial imaging systems, where high fidelity images are required for efficient and reliable real-time quantitative analysis of the acquired image. Moreover, the project has further strengthened the cooperation between the University of Ljubljana, Jožef Stefan Institute and Technology Park Ljubljana, in which the co-financing company Sensum, Computer Vision Systems resides. This scientific and technological cooperation has already led to the development of numerous high-tech computer vision products that are used for quality inspection and sorting of tablets and capsules in the pharmaceutical companies worldwide (see www.sensum.eu). The applied project further enhanced and extended the existing cooperation between the R&D institutions and enterprises and, thereby, significantly contributed to the transfer of top scientific knowledge to industry, which is essential for development of new high-tech products with high value added, accelerating the economic growth, and generating new well-paid jobs.
Most important scientific results Annual report 2011, 2012, 2013, final report, complete report on dLib.si
Most important socioeconomically and culturally relevant results Annual report 2012, 2013, final report, complete report on dLib.si
Views history