EDI – Elektronikas un datorzinātņu institūts
Back to staff

Mārtiņš Puķītis

Mg. math.

Research Assistant

Mārtiņš Puķītis is a research assistant at the Institute of Electronics and Computer Science (EDI), where he began working in 2015. In the same year, he obtained a master's degree in mathematics from the Faculty of Physics and Mathematics at the University of Latvia.

Mārtiņš research interests are related to remote sensing, with a particular focus on the automatic classification of images obtained by remote sensing. His work involves the application of various image processing, machine learning, and data analysis methods to improve image interpretation and ensure effective information extraction from satellite and other image data.

Projects

1 result

Publications

  • 2023

    Classification of satellite images using Dynland technology(opens in a new window)

    Martins Pukitis, Ints Mednieks

    Ninth International Conference on Remote Sensing and Geoinformation of the Environment (RSCy2023), Vol. 127860, Proc. SPIE 12786

    Abstract

    Satellite images are widely used for land cover classification into predefined categories. To perform it in a practicable way, field data are needed to train and test a classifier. The collection of accurate field data requires significant resources, and image data of a category may have sophisticated distribution in the feature space. “Dynland” technology was developed to overcome these problems and offer a robust and practical way to classify such images. It is based on a nonparametric clustering method applicable to data with different distributions with a following assignment of classes to clusters based on their overlapping with available reference data which can be scarce and imprecise. Dynland clustering algorithm provides meaningful clusters but requires huge computational resources currently limiting the size of processed images. To overcome this problem, we propose to apply a sampling procedure i.e. use each 𝑛-th pixel on both axes for clustering and distribute the remaining pixels to created clusters based on nearest neighbors’ search. However, such an approach reduces the classification accuracy; the analysis results of this reduction will be presented. Another way how to deal with images of large areas is to split them into fragments. To classify a fragment, reference data for all categories of interest should be available for the area it represents. This limitation should be resolved so that the available reference data are used for areas not limited to one fragment. To deal with that, a method is proposed allowing to exploit the classification result of a neighboring fragment for which the reference data were available in full. The proposed solutions widened the applicability of the Dynland technology e.g. enabled the production of a classified map of mires and peatlands for the whole territory of Latvia.

    Indexed in Scopus / WoS

  • 2023

    Wetland Change Detection Using Sentinel-2 in the Part of Latvia(opens in a new window)

    Andris Skromulis, Juris Breidaks, Mārtiņš Puķītis

    Environment. Technology. Resources. Proceedings of the International Scientific and Practical Conference June 15-16, 2023, Vol. 1, Rezekne Academy of Technologies

    Abstract

    In the article, the possible impact of changes on wetland were analysed by the semi-supervised classification method of statistical analysis. The Sentinel-2 raw data between two different seasons are combined together. The data preparation is shortly described in the article. Data is clustered with unsupervised method. The article describes a supervised method – how data credibility and classification can be estimated if its reference is poor quality.

    Indexed in Scopus / WoS

  • 2016

    Fusion of Multisensor Data Based on Different Multidimensional Distributions

    LORENCS, A., MEDNIEKS, I., SINICA-SINAVSKIS, J., PUKITIS M.

    Abstract

    <p>LORENCS, A., MEDNIEKS, I., SINICA-SINAVSKIS, J., PUKITIS M. Fusion of Multisensor Data Based on Different Multidimensional Distributions, Electronics and Electrical Engineering (Elektronika ir Elektrotechnika, Vol. 22, No.4, pp. 68-72, 2016.<br /> Abstract</p> <p>The paper addresses problems related with classification of images obtained using different types of remote sensors. Development and use of Bayes type land cover classifiers based on multidimensional Gaussian, Dirichlet and gamma distributions is analysed and compared on the basis of sample data from RGB and hyperspectral thermal sensors with unequal spatial resolution. Approaches to data fusion for design of the combined classifiers are presented including the cases where different families of multidimensional distributions are used to model the sensor data and classifiers are designed using combinations of their probability density functions. The best classification results are obtained when the fusion of data from both images is used together with classification based on all three considered distributions combined together.</p>