Daniel Sawicki

FLAME MONITORING USING IMAGE CLASSIFICATION

This paper presents comparison of image classification methods for co-firing biomass and pulverized coal. Two classes of combustion - stable and unstable were defined for nine variants with different power value parameters and fixed amount biomass. Experimental results show that correct classification of images was achieved for the assumed variants.

Evaluation of uncertainty of control by measurement with logical conditions

The method for evaluation of uncertainty in control by measurement systems with logical conditions under impact of parametric perturbations was proposed. The method is based on the linearization relay and logical transformations. Relay ones were linearized by harmonic method. Logical ones were linearized by arithmetic decomposition of logical function. Efficiency of method proved by comparison of results in simulation system.

Coloring VBR Streams Inside the DiffServ Domain

Video traffic is supposed to account for a large portion of future wired and wireless network traffic. The evaluation of different video coding standards for their effects on the network traffic, and the resulting requirements for networks have attracted great interest in the research community. Simultaneously, the numerous test and users’ experiences proved that the limited quality of service (QoS) features available in both IPV4 and IPv6 cannot accommodate the various degrees of requirements needed by multimedia traffic and in particular, the video traffic. This paper investigates the influence of various packet markers on VBR video streams inside the DiffServ domain. Simulations based on NS-2 network simulator and Evalvid-RA framework and follow the IETF recommendations for video traffic shaping in the IP networks.

Determining of combustion process state based on flame images analysis using k-NN classification

This paper presents comparison image classification method of combustion biomass and pulverized coal. Presented research is related with 20% weight fraction of the biomass. Defined two class of combustion: stable and unstable for nine variants with different power, secondary air value parameters and fixed amount biomass. Used k-nearest neighbors algorithm classification to test, validation and classify flame image which correspond with the state of the combustion process.

Computer-aided system for interactive psychomotor testing

Nowadays research of psychomotor actions has taken a special place in education, sports, medicine, psychology etc. Development of computer system for psychomotor testing could help solve many operational problems in psychoneurology and psychophysiology and also determine the individual characteristics of fine motor skills. This is particularly relevant issue when it comes to children, students, athletes for definition of personal and professional features. The article presents the dynamics of a developing psychomotor skills and application in the training process of means. The results of testing indicated their significant impact on psychomotor skills development.

The use of combined classification for analysis of the combustion process state

In present work comparison combined image classification method of co-firing biomass and pulverized coal are proposed. The images were captured by vision monitoring system with camera and a borescope. Presented research is related with 20% weight fraction of the biomass. Defined two class of combustion: stable and unstable for nine variants with different power, secondary air value parameters and fixed amount biomass. Used combined classification of algorithm (MLP, SVM, k-NN, LDA, QDA) to classify flame image which correspond with the state of the combustion process.

Determining of combustion process state based on flame images analysis

This paper presents comparison image classification method of combustion biomass and pulverized coal. Presented research is related with 10% and 20% weight fraction of the biomass. Defined two class of combustion: stable and unstable for nine variants with different power, secondary air value parameters and fixed amount biomass. Used artificial neural networks and support vector machine to classify flame image which correspond with the state of the. combustion process.

Monitoring combustion process with the vision diagnostic system

This paper presents comparison image classification method of co-firing biomass and pulverized coal. Defined two class of combustion: stable and unstable for three variants with different power value parameters and fixed amount biomass. Used support vector machine to classify flame image which correspond with the state of the combustion process.

Using PCA in flame vision monitoring system

This paper presents comparison image classification method of co-firing biomass and pulverized coal. Defined two class of combustion: stable and unstable for three variants with different power value parameters and fixed amount biomass. Used Principal component analysis for determine the most important features that affect the state of the combustion process.