Grupo de Ingeniería Estadística Multivariante GIEM

Loading...
OrgUnit Logo
Date established
City
Country
ID
Description

Publication Search Results

Now showing 1 - 3 of 3
  • Publication
    PLS: A versatile tool for industrial process improvement and optimization
    (John Wiley & Sons, 2008-12) Ferrer Riquelme, Alberto José; Aguado García, Daniel; Vidal Puig, Santiago; Prats Montalbán, José Manuel; Zarzo Castelló, Manuel; Dpto. de Ingeniería Hidráulica y Medio Ambiente; Dpto. de Estadística e Investigación Operativa Aplicadas y Calidad; Instituto Universitario de Ingeniería del Agua y del Medio Ambiente; Escuela Técnica Superior de Ingeniería de Caminos, Canales y Puertos; Escuela Técnica Superior de Ingeniería Industrial; Escuela Técnica Superior de Ingeniería Informática; Escuela Técnica Superior de Ingeniería Agronómica y del Medio Natural; Grupo de Ingeniería Estadística Multivariante GIEM; European Commission; Ministerio de Ciencia y Tecnología
    [EN] Modern industrial processes are characterized by acquiring massive amounts of highly collinear data. In this context, partial least-squares (PLS) regression, if wisely used, can become a strategic tool for process improvement and optimization. In this paper we illustrate the versatility of this technique through several real case studies that basically differ in the structure of the X matrix (process variables) and Y matrix (response parameters). By using the PLS approach, the results show that it is possible to build predictive models (soft sensors) for monitoring the performance of a wastewater treatment plant, to help in the diagnosis of a complex batch polymerization process, to develop in automatic classifier based on image data, or to assist in the empirical model building of a continuous polymerization process.
  • Publication
    Comparison of latent variable-based and artificial intelligence methods for impurity detection in PET recycling from NIR hyperspectral images
    (John Wiley & Sons, 2018) Galdón-Navarro, Borja; Prats Montalbán, José Manuel; Cubero-García, Sergio; Blasco Ivars, Jose; Ferrer Riquelme, Alberto José; Dpto. de Estadística e Investigación Operativa Aplicadas y Calidad; Escuela Técnica Superior de Ingeniería Industrial; Grupo de Ingeniería Estadística Multivariante GIEM; Universitat Politècnica de València
    [EN] In polyethylene terephthalate's (PET)'s recycling processes, separation from polyvinyl chloride (PVC) is of prior relevance due to its toxicity, which degrades the final quality of recycled PET. Moreover, the potential presence of some polymers in mixed plastics (such as PVC in PET) is a key aspect for the use of recycled plastic in products such as medical equipment, toys, or food packaging. Many works have dealt with plastic classification by hyperspectral imaging, although only some of them have been directly focused on PET sorting and very few on its separation from PVC. These works use different classification models and preprocessing techniques and show their performance for the problem at hand. However, still, there is a lack of methodology to address the goal of comparing and finding the best model and preprocessing technique. Thus, this paper presents a design of experiments-based methodology for comparing and selecting, for the problem at hand, the best preprocessing technique, and the best latent variable-based and/or artificial intelligence classification method, when using NIR hyperspectral images. There is a lack of methodology to address the goal of comparing and finding the best model and preprocessing technique. Thus, this paper presents a design of experiments-based methodology for comparing and selecting, for the problem at hand, the best preprocessing technique, and the best latent variable-based and/or artificial intelligence classification method when using near-infrared hyperspectral images.
  • Publication
    Pixel classification methods for identifying and quantifying leaf surface injury from digital images
    (Elsevier, 2014-10) Opstad Kruse, Ole Mathis; Prats Montalbán, José Manuel; Indahl, Ulf Geir; Kvaal, Knut; Ferrer Riquelme, Alberto José; Futsaether, Cecilia Marie; Dpto. de Estadística e Investigación Operativa Aplicadas y Calidad; Escuela Técnica Superior de Ingeniería Industrial; Grupo de Ingeniería Estadística Multivariante GIEM
    Plants exposed to stress due to pollution, disease or nutrient deficiency often develop visible symptoms on leaves such as spots, colour changes and necrotic regions. Early symptom detection is important for precision agriculture, environmental monitoring using bio-indicators and quality assessment of leafy vegetables. Leaf injury is usually assessed by visual inspection, which is labour-intensive and to a consid- erable extent subjective. In this study, methods for classifying individual pixels as healthy or injured from images of clover leaves exposed to the air pollutant ozone were tested and compared. RGB images of the leaves were acquired under controlled conditions in a laboratory using a standard digital SLR camera. Different feature vectors were extracted from the images by including different colour and texture (spa- tial) information. Four approaches to classification were evaluated: (1) Fit to a Pattern Multivariate Image Analysis (FPM) combined with T2 statistics (FPM-T2) or (2) Residual Sum of Squares statistics (FPM-RSS), (3) linear discriminant analysis (LDA) and (4) K-means clustering. The predicted leaf pixel classifications were trained from and compared to manually segmented images to evaluate classification performance. The LDA classifier outperformed the three other approaches in pixel identification with significantly higher accuracy, precision, true positive rate and F-score and significantly lower false positive rate and computation time. A feature vector of single pixel colour channel intensities was sufficient for capturing the information relevant for pixel identification. Including neighbourhood pixel information in the feature vector did not improve performance, but significantly increased the computation time. The LDA classifier was robust with 95% mean accuracy, 83% mean true positive rate and 2% mean false positive rate, indicating that it has potential for real-time applications.