Dalibor Šeljmeši

Mr. Dalibor Šeljmeši

University in Novi Sad Technical Faculty "Mihajlo Pupin", Zrenjanin, Serbia
orcid

Research Focus & Metrics

Academic Career Length

3Years (2024-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
4 July 2024 in Applied Sciences
Verified
Journal Article
Open access
PDF file
LJLjubisa Josimović
Ljubisa Josimović
SPSlavica Prvulovic
Slavica Prvulovic
Luka DjordjevićLuka Djordjević
Luka Djordjević
IBIvana Bicok
Ivana Bicok
MBMihalj Bakator
Mihalj Bakator
DŠDalibor Šeljmeši
Dalibor Šeljmeši
This paper aims to enhance the efficiency of biogas plants for the production of electrical and thermal energy by optimizing substrate selection and digester heating techniques. The study presents a comprehensive project of a biogas plant, including all necessary installations and a detailed methodology for equipment selection. Specific substrates used include animal flour, corn silage, and molasses, each chosen for their high dry matter content and biogas production efficiency. The digester heating technique was analyzed to maintain optimal fermentation temperatures, ensuring a stable and efficient biogas production process. The projected daily biogas production is estimated to be 5688 m3. The analysis shows that maintaining a constant digester temperature significantly improves biogas yield. Seasonal variations in thermal energy requirements were identified, highlighting the need for effective insulation and heating mechanisms, particularly during colder months. Additionally, the implementation of advanced heating techniques demonstrated a reduction in overall energy consumption for maintaining the desired digester temperature. This research provides valuable insights for the design and operation of efficient biogas plants. By optimizing substrate mixtures and digester heating, the study contributes to sustainable energy production, reduced greenhouse gas emissions, and enhanced biogas plant efficiency, supporting both environmental protection and economic development.

15 April 2026 in AI
Verified
Journal Article
Open access
PDF file
DŠDalibor Šeljmeši
Dalibor Šeljmeši
VBVladimir Brtka
Vladimir Brtka
VIVelibor Ilic
Velibor Ilic
Dalibor DobrilovicDalibor Dobrilovic
Dalibor Dobrilovic
EBEleonora Brtka
Eleonora Brtka
VOVišnja Ognjenović
Višnja Ognjenović
Reliable change detection in indoor environments remains a challenge for autonomous robotic systems using 3D LiDAR. Existing methods often require manual annotation, computationally intensive architectures, or focus on outdoor scenes. This paper presents SPICD-Net, a lightweight Siamese PointNet framework for indoor 3D change detection trained exclusively on synthetically generated anomalies, eliminating manual labeling. The framework offers three deployment-oriented contributions: a three-class Siamese formulation separating no-change, changed, and geometrically inconsistent tile pairs; a pre-FPS anomaly injection strategy that aligns synthetic training with inference-time preprocessing; and a stochastic-gated Chamfer-statistics branch that complements learned embeddings with explicit geometric cues under consumer-grade hardware constraints. Evaluated on 14 controlled simulation experiments in an indoor corridor dataset, SPICD-Net achieved aggregated Precision = 0.86, Recall = 0.82, F1-score = 0.84, and Accuracy = 0.96, with zero false positives in the no-change baseline and mean inference time of 22.4 s for a 172-tile map on a single consumer GPU. Additional robustness experiments identified registration accuracy as the main operational prerequisite. A limited real-world validation in one unseen room (four scans, 67 tiles) achieved Precision = 0.583, Recall = 1.000, and F1 = 0.737.

1 January 2025
Verified
Conference Paper
Closed access
DŠDalibor Šeljmeši
Dalibor Šeljmeši
VOVišnja Ognjenović
Višnja Ognjenović
VIVelibor Ilic
Velibor Ilic
Dalibor DobrilovicDalibor Dobrilovic
Dalibor Dobrilovic
VBVladimir Brtka
Vladimir Brtka
Milan NikolićMilan Nikolić
Milan Nikolić
This paper presents an automated system for constructing global 3D maps from mobile LiDAR scans collected using a Unitree Go2 quadruped robot. The proposed offline pipeline includes overlap detection, point-to-plane ICP registration with quality assessment, pose graph optimization, and generation of a spatially indexed global map. Evaluated on real industrial data, the system produces a stable main map component despite limited overlap and geometrically uniform environments, achieving an average alignment RMSE of 2.5 cm. The resulting map is suitable for downstream 3D machine-learning tasks such as change detection and forms a foundation for digital twin applications in Industry 4.0/5.0.