The energy emitted by a flame and its spectra provide critical information about fuels, such as energy type, temperature, and molecular properties. While traditionally used to analyze combustion, we introduce a spectral-time-based index, optical calorific power (OCP), to calculate the total energy released per mass unit (J/Kg), relating to the calorific value of wood-based fuels. Using novel optical variables and machine learning, we identified different wood species during the combustion of the wood samples. Homogeneous samples of four wood species (sapwood and heartwood) were ignited in a temperature-controlled furnace. Spectra were measured using a calibrated spectrophotometer in the 450-900 nm range with a 100ms integration time. Continuous and discontinuous spectral patterns were observed in all samples, used to calculate spectral-time-based optical variables, and separated using the AirPLS algorithm. Discontinuous spectra correlated with Na and K emissions (589.4 nm and 766.5 nm, respectively). Continuous spectra were analyzed to determine optical variables such as flame temperature (K), total continuous radiation (TCR, mu W/cm(2)), and total continuous energy (TCE, mu J). Five supervised classification models, XGBoost, LR, SVM, LDA, and RF, were trained using normalized physical parameters and optical variables with stratified cross-validation. The proposed optical variables yielded an identification accuracy of 93%, precision of 95%, and recall of 93% during combustion using the XGBoost method. These promising results demonstrate the potential of our method for accurately identifying wood species, while offering a cost-effective and novel alternative to NIR-SWIR reflectance spectral identification techniques.

The correct identification of minerals is crucial task for the exploration and exploitation of mineral resources, environmental monitoring, and industrial processes. In this article, we propose a hyperspectral imaging system and classification model to identify nine types of minerals. To accomplish this, we employed a hyperspectral shortwave infrared (SWIR) camera to capture hyperspectral images. We then introduce a convolutional neural network (CNN) architecture that considers only spectral data, complemented by a fully connected network for classification. To prevent overfitting, we implemented the dropout technique, which enables random deactivation of neurons during the backpropagation process. This results in improved performance during the training phase and a better generalization capacity. Training was optimized to minimize the categorical cross-entropy objective function, and the model was evaluated during training using an accuracy metric. Finally, we evaluated the results with the test data using accuracy, recall, and precision metrics, and achieved 98.52%, 98.25%, and 98.68%, respectively. Our source code is available at https://github.com/jcifuenr/Spec-CNN.

Flame spectroscopy is a technique widely used to analyze combustion processes.In a flame, energy is emitted across a wide spectral range, and its corresponding spectra can be classified in both, continuous and discontinuous behaviors. Optimizing combustion, has the potential to maximize efficiency, thus reducing fuel consumption and emissions of residual gases. This report shows the spectral emission of a flame emitted by combustion of pellets at different humidity contents, from a specific brand (here referred to as Alfa), typical in the Chilean market. To perform the spectral analysis, the HR-4000 spectrophotometer (Ocean Optics), previously calibrated in absolute radiance (in uW/nm·cm 2 ) was used. From the collected spectra, parameters such as flame temperature (in °C), total continuous radiation (in µW/cm2) and total continuous energy (in µJ) were estimated. The spectral behavior, at different humidity contents, shown different spectral patterns. Thus, in this report we introduce the optical calorific power (in J/kg), quantifying the energy provided by a pellet sample, and calculated from spectral results. For the Alfa brand, with a humidity content between a 6-8% we estimated an optical calorific power of 6.12 (J/kg) and with a humidity content of 13.6 %, an optical calorific power of 4.96 (J/kg). The different parameters proposed in this work hold promises for understanding biomass combustion phenomena, such as the dynamics of energy emission in flame distribution.

A flame emits energy over a wide spectral region, providing important information about the combustion process. Its associated spectra, classified into continuous and discontinuous features, provide information concerning the temperature of the soot particle and atomic or molecular reactions. However, because of the non-linear and turbulent nature of the flame, added to the high-dimensional nature of the spectral data and the overlap between continuous and discontinuous spectral emissions, complex methods are needed to obtain valuable information for combustion closed-loop control and optimization. In this paper, we propose a method for retrieval of hyperspectral flame images from Liquefied Petroleum Gas (LPG) fuel, using only images captured by a trichromatic camera, but combined with a spectral recovery method. The hyperspectral flame data recovered was used to estimate indexes, such as the flame temperature (in K), the local emitted radiance (at the pixel level) and the global radiance (both in μW/cm2), thus performing a dynamic, spectral and spatial combustion diagnosis in different air/fuel ratios. Our method was validated by performing measurements with a calibrated spectrophotometer, achieving small errors in the retrieval procedure (less than 1%) and in the temperature estimation (less than 4%). The results show that our approach can play an essential role in flame sensing, providing an important tool for monitoring or controlling combustion processes.

A method is proposed to validate the flame temperature calculated with the two-color pyrometer method. The temperature is obtained from the estimated spectra, us- ing the Maloney & Wandell method, and compared with measurements from a type K thermocouple.

A flame emits energy over a wide spectral region, and its associated spectra contain both continuous and discontinuous components, closely related to the combustion process current state. However, the non-linear and turbulent nature of the flame, high dimensionality of the spectral data, and the overlap between continuous and discontinuous spectral emissions requires the use of complex method to obtain valuable information for combustion closed loop-control and optimization. This work analyzes the spectral emission of the flame emitted by LPG flames to compute high resolution spectra combining the Maloney-Wandell spectral recovery method and field information of flame images with low spectral resolution. Then, we obtain different variables such as the flame temperature (in K) and local radiance (in μW=cm2), performing a dynamic, spectral and spatial combustion diagnostics.

Despite the widespread and extended use of sensor systems in different industries, there is an important gap to reach equivalent conditions in pyrometallurgical processes for primary production. In the specific case of copper pyrometallurgy, the situation is particularly challenging to incorporate the Industry 4.0 concept for the optimization of their operations. Currently, only two instruments can be identified at the commercial level: the Noranda pyrometer and the Online Production Control (OPC) system. The iron-making and steelmaking industries, however, present an advanced level of control based on monitoring and sensing networks throughout the entire process. This reality has served as a basis for developing a series of solutions based on radiometric sensors for copper pyrometallurgy. We present two types of sensing concept. The first one is applied to smelting and converting reactors based on the measurements of the radiation of the oxidation of different copper and iron sulfides. The second one considers hyperspectral imaging of molten phases flow during operations. The idea of this proposal is to transfer some commercial sensing technologies already in use in the steelmaking industry. In this article, the fundamentals of the sensor design, proofs of concept, and the initial industrial validations are reviewed. Finally, a discussion on the contribution of this knowledge and development opportunities within the framework of Industry 4.0 are addressed.

In process industries, the availability of large volumes of data is not directly related to the extraction of valuable information or process monitoring with good performance. Usually, data is directly visualized as tables or tendency graphics, being not used properly. This paper presents the design of process monitoring by considering the design of alarms visualization plots which provides useful information in a unique plot, combined with soft sensor design used for the prediction of critical variables which defines the process operational performance. Examples of two cases with real industrial data are provided to demonstrate the effectiveness and utility of these methods.

The project focused on the investigation of new sensing techniques in the copper production industry, specifically in the flash smelting of copper concentrate process. In this paper, we report a direct relationship between the visible and near-infrared emission spectra in the combustion of copper concentrates by changing some operating conditions such as the sulfur-copper and the oxygen ratios provided to the reaction zone. Spectral processing techniques are applied to the measured spectra. The first one aims to separate both continuous radiations mainly associated to incandescent particles and heating walls with discontinuous emissions associated with some emitting atoms and molecules. This goal was carried out using airPLS baseline estimation algorithm. The second processing technique aims to find the continuous emission only associated with the combustion of copper concentrate particles, eliminating the background spectra associated with the smelter walls. This goal was carried out by directly measuring walls emission at operating temperature and in the absence of flame. The most relevant results show that the estimation of the total radiation associated with each measured spectra is an intrinsic parameter of the process that can provide useful information to the operator that supervises the industrial process. It allows estimate quantitatively the sulfur-copper ratio in order to online monitor the mineral characteristics of the copper concentrate that is entering the process. On the other hand, the approach of a first prototype of the emissivity model for the copper concentrate particles, which validated with measurements, becomes a promising tool that will allow increasing the development of optoelectronic applications around this industry process.

Combustion is at the core of many of industrial processes where heat transfer occurs. The standard setup controls the flame behavior by measuring the composition of the residual gases discharged to the atmosphere. However, such a solution involves an unavoidable transport delay due to the distance between the flame and the measurement sample, in addition to intrinsic delay in the sensor. In this work, we propose the optically monitoring of the combustion inferring the state of the process by the inclusion of a spectrometer to calculates a novel set of optical variables. Such approach provides an improvement in the monitoring and control of the process by calculating an accurate estimation of the energy efficiency by measuring the flame temperature Tf, the total radiation Radt proportional to the heat transfer; or by measuring the radicals ratio C2*/CH*. We then present the fundamentals of a novel proposition for control and optimization of industrial combustion processes, based on the use of flame emissions information measured by the spectrometer.

A proper mechatronic system capable of neutralizing people entering unauthorized areas by firing a paintball gun is presented in this paper. The effective shooting range is 25 meters limited only by the range of the marker. The implemented system is capable of operating in different light conditions, at a maximum distance of 70 meters. The mechanism, together with the artificial vision system move in two axes: turn and tilt. Processing includes identification, tracking, pan tilt zooms control and automatic trigger activation control. The tests were carried out in a stadium, following the safety rules to avoid damage to the test subject. The shooting effectiveness is 80%.

In this paper, we present a method for hyperspectral retrieval using multispectral satellite images. The method consists of the use of training spectral data with a compressive capability. By using principal component analysis (PCA), a proper number of basis vectors are extracted. These vectors are properly combined and weighted by the sensors’ responses from visible MODIS channels, achieving as a result the retrieval of hyperspectral images. Once MODIS channels are used for hyperspectral retrieval, the training spectra are projected over the recovered data, and the ground-based process used for training can be reliably detected. To probe the method, we use only four visible images from MODIS for large-scale ash clouds’ monitoring from volcanic eruptions. A high-spectral resolution data of reflectances from ash was measured in the laboratory. Using PCA, we select four basis vectors, which combined with MODIS sensors responses, allows estimating hyperspectral images. By comparing both the estimated hyperspectral images and the training spectra, it is feasible to identify the presence of ash clouds at a pixel-by-pixel level, even in the presence of water clouds. Finally, by using a radiometric model applied over hyperspectral retrieved data, the relative concentration of the volcanic ash in the cloud is obtained. The performance of the proposed method is compared with the classical method based on temperature differences (using infrared MODIS channels), and the results show an excellent match, outperforming the infrared-based approach. This proposal opens new avenues to increase the potential of multispectral remote systems, which can be even extended to other applications and spectral bands for remote sensing. The results show that the method could play an essential role by providing more accurate information of volcanic ash spatial dispersion, enabling one to prevent several hazards related to volcanic ash where volcanoes’ monitoring is not feasible.

This paper presents the results of two novel approaches to measure the soot propensity of a flame and their comparison with the Line of Sight Attenuation (LOSA) method. Both approaches are based on the detection of the Smoke Point Height (SPH), concept used to determine when a flame is in a sooting state. The first approach is based on the detection of morphological changes in the flame, identified through their amplification via the Eulerian Video Magnification algorithm. Results show an effective amplification of the flame geometry, allowing the visualization of variations on the flame tip unable to be detected by the naked human eye and therefore the detection of SPH. The second approach is based on the application of Artificial Intelligence models to classify flame images regarding their sooting propensity, taking advantage of the knowledge acquired from a referential data set. Both approaches provide an accurate classification when compared to the conventional method of LOSA. Furthermore, both approaches show a greater implementation potential in practical combustion devices than the conventional method of LOSA, due to their reduced hardware and technical requirements.

The present project implements an application for the search, recognition and monitoring of people based on artificial vision algorithms. The OpenCV libraries are used to process the images, which were obtained from a conventional IP video surveillance camera. This type of cameras can be used in different environmental conditions (high, medium and low lighting) and up to an effective distance of 70 m. In the detection and search phase, cascade classifiers are used with local binary patterns LBP (Local Binary Patterns). Subsequently, in the follow-up phase, a tracking algorithm is implemented, addressed only to the person detected through kernelized correlation filters KCF (Kernelized Correlation Filters), so that the objective is not lost. A graphical interface was developed in the Qt Software which allows an easy use of the application. The average effectiveness of the algorithm is 90

Visible spectral emissions emitted by different biomass species in a combustion process were measured and analyzed. By applying spectral techniques, it is possible to characterize and to monitor the combustion process, identifying different species.

In this paper, we report on spectral features emitted by a reaction shaft occurring in flash smelting of copper concentrates containing sulfide copper minerals such as chalcopyrite (CuFeS2), bornite (Cu5FeS4) and pyrite (FeS2). Different combustion conditions are addressed, such as sulfur-copper ratio and oxygen excess. Temperature and spectral emissivity features are estimated for each case by using the two wavelength method and radiometric models. The most relevant results have shown an increasing intensity behavior for higher sulfur-copper ratios and oxygen contents, where emissivity is almost constant along the visible spectrum range for all cases, which validates the gray body assumption. CuO and FeO emission line features along the visible spectrum appear to be a sensing alternative for describing the combustion reactions.

Optimization of combustion processes holds the promise of maximizing energy efficiency, at the same time lowering fuel consumption and residual gases emissions. In this context, the current common operation setting in combustion processes could be improved by the introduction of passive optical sensors, which can be located close the flame, thus eliminating the inherent transport delay in current setups that only infer the combustion quality by measuring residual gases emissions. However, there is a tradeoff for flame detection between spatial-spectral resolutions, depending on the optical sensor scheme. In this paper, we present the fundamentals to avoid this constraint, obtaining a combined high spectral and spatial resolution measurement suitable for combustion diagnostics and control. The core of this proposal is to use the flame images from a low-spectral resolution charge-coupled device camera, combined with a spectral recovery method. This method is based on the off-line samples measured on the continuous component of flame spectra, providing a set of vector basis to estimate a calibrated flame spectra at each pixel. The results of the spectral recovery process verify the suitability of the method in terms of goodness-of-fit coefficient and root mean square error metrics, enabling hyper-spectral measurements based on the combination of different optical sensors. Then, continuous estimated spectra along the flame are used to calculate the energy transfer released by radiation, useful for combustion diagnostics.

In this work, we set the foundations for the design of a control system based on the measurement of flame's total radiation from a combustion process. The key aspect of this approach is that flame radiation from the combustion process can be successfully modeled by means of a Hammerstein system. The static nonlinearity, present in the examples studied by the authors, proves to be mild and measurable in absolute terms using a flame spectral analysis based method. As a real life example, we describe the closed-loop solution applied for the case of a ladle furnace preheating process. Our results show that the proposed controller is effectively superior to the existing open loop solution commonly used in industry.

In this work we analyze the spectral emission of the flame emitted by burning biomass, of different species of wood such as Cypress, Fig, Olive and Pine. The analysis was carried out to distinguish the different species of wood sensing the compounds of Sodium and Potassium emitted at 582 nm and 780 nm. Also, a spectral analysis carried out for samples of wood at different contents of water. The results show that is possible make a difference between the biomass species, and a new technique can be implemented that will allow the estimation of water content in wood using the emission spectrum of the burned wood.