WO2016061647A1 - Processo de análise de particulados via modelagem por processamento de imagens digitais - Google Patents
Processo de análise de particulados via modelagem por processamento de imagens digitais Download PDFInfo
- Publication number
- WO2016061647A1 WO2016061647A1 PCT/BR2014/000389 BR2014000389W WO2016061647A1 WO 2016061647 A1 WO2016061647 A1 WO 2016061647A1 BR 2014000389 W BR2014000389 W BR 2014000389W WO 2016061647 A1 WO2016061647 A1 WO 2016061647A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- modeling
- analysis
- sub
- particulate
- image processing
- Prior art date
Links
- 238000000034 method Methods 0.000 title claims abstract description 112
- 238000012545 processing Methods 0.000 title claims abstract description 53
- 238000004458 analytical method Methods 0.000 claims abstract description 57
- 238000001228 spectrum Methods 0.000 claims abstract description 34
- 230000008569 process Effects 0.000 claims description 86
- 239000002245 particle Substances 0.000 claims description 25
- 238000013075 data extraction Methods 0.000 claims description 13
- 238000003384 imaging method Methods 0.000 claims description 13
- 238000007596 consolidation process Methods 0.000 claims description 10
- 238000007705 chemical test Methods 0.000 claims description 8
- 239000011159 matrix material Substances 0.000 claims description 8
- 238000002513 implantation Methods 0.000 claims description 7
- 229910052500 inorganic mineral Inorganic materials 0.000 claims description 7
- 239000011707 mineral Substances 0.000 claims description 7
- 238000013098 chemical test method Methods 0.000 claims description 5
- 238000013031 physical testing Methods 0.000 claims description 5
- 238000005070 sampling Methods 0.000 claims description 5
- 238000000605 extraction Methods 0.000 claims description 4
- 238000005286 illumination Methods 0.000 claims description 4
- 239000013618 particulate matter Substances 0.000 claims description 4
- 238000002360 preparation method Methods 0.000 claims description 4
- 238000011002 quantification Methods 0.000 claims description 4
- 238000013528 artificial neural network Methods 0.000 claims description 3
- 230000003287 optical effect Effects 0.000 claims description 3
- 230000001174 ascending effect Effects 0.000 claims description 2
- 238000005094 computer simulation Methods 0.000 claims description 2
- 238000012797 qualification Methods 0.000 claims description 2
- 238000000926 separation method Methods 0.000 claims description 2
- 239000011236 particulate material Substances 0.000 abstract description 4
- 238000013480 data collection Methods 0.000 abstract 1
- 238000004904 shortening Methods 0.000 abstract 1
- VYPSYNLAJGMNEJ-UHFFFAOYSA-N Silicium dioxide Chemical compound O=[Si]=O VYPSYNLAJGMNEJ-UHFFFAOYSA-N 0.000 description 58
- XEEYBQQBJWHFJM-UHFFFAOYSA-N Iron Chemical compound [Fe] XEEYBQQBJWHFJM-UHFFFAOYSA-N 0.000 description 56
- 239000000377 silicon dioxide Substances 0.000 description 29
- 229910052742 iron Inorganic materials 0.000 description 28
- 239000000126 substance Substances 0.000 description 19
- 239000003086 colorant Substances 0.000 description 11
- 241000894007 species Species 0.000 description 10
- 239000000203 mixture Substances 0.000 description 9
- 238000004519 manufacturing process Methods 0.000 description 8
- 238000012360 testing method Methods 0.000 description 8
- 238000010200 validation analysis Methods 0.000 description 8
- 239000008188 pellet Substances 0.000 description 7
- 238000005065 mining Methods 0.000 description 6
- 238000003491 array Methods 0.000 description 5
- 238000005188 flotation Methods 0.000 description 5
- 229910001608 iron mineral Inorganic materials 0.000 description 5
- 238000005259 measurement Methods 0.000 description 5
- 230000008901 benefit Effects 0.000 description 4
- 239000000284 extract Substances 0.000 description 4
- 230000008447 perception Effects 0.000 description 4
- 230000009466 transformation Effects 0.000 description 4
- 238000010276 construction Methods 0.000 description 3
- 238000001033 granulometry Methods 0.000 description 3
- 239000002994 raw material Substances 0.000 description 3
- 230000011218 segmentation Effects 0.000 description 3
- 210000004556 brain Anatomy 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 238000009826 distribution Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 230000007613 environmental effect Effects 0.000 description 2
- 230000004438 eyesight Effects 0.000 description 2
- 230000000574 ganglionic effect Effects 0.000 description 2
- 229910052598 goethite Inorganic materials 0.000 description 2
- 238000000227 grinding Methods 0.000 description 2
- 239000011019 hematite Substances 0.000 description 2
- 229910052595 hematite Inorganic materials 0.000 description 2
- AEIXRCIKZIZYPM-UHFFFAOYSA-M hydroxy(oxo)iron Chemical compound [O][Fe]O AEIXRCIKZIZYPM-UHFFFAOYSA-M 0.000 description 2
- 238000003709 image segmentation Methods 0.000 description 2
- 230000003993 interaction Effects 0.000 description 2
- LIKBJVNGSGBSGK-UHFFFAOYSA-N iron(3+);oxygen(2-) Chemical compound [O-2].[O-2].[O-2].[Fe+3].[Fe+3] LIKBJVNGSGBSGK-UHFFFAOYSA-N 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 230000005855 radiation Effects 0.000 description 2
- 230000002207 retinal effect Effects 0.000 description 2
- 230000003595 spectral effect Effects 0.000 description 2
- 238000000844 transformation Methods 0.000 description 2
- WKBPZYKAUNRMKP-UHFFFAOYSA-N 1-[2-(2,4-dichlorophenyl)pentyl]1,2,4-triazole Chemical compound C=1C=C(Cl)C=C(Cl)C=1C(CCC)CN1C=NC=N1 WKBPZYKAUNRMKP-UHFFFAOYSA-N 0.000 description 1
- 101100186820 Drosophila melanogaster sicily gene Proteins 0.000 description 1
- CWYNVVGOOAEACU-UHFFFAOYSA-N Fe2+ Chemical compound [Fe+2] CWYNVVGOOAEACU-UHFFFAOYSA-N 0.000 description 1
- 229920002472 Starch Polymers 0.000 description 1
- 229910000831 Steel Inorganic materials 0.000 description 1
- 230000001464 adherent effect Effects 0.000 description 1
- 239000010975 amethyst Substances 0.000 description 1
- 150000001412 amines Chemical class 0.000 description 1
- 239000012620 biological material Substances 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 239000013043 chemical agent Substances 0.000 description 1
- 238000009614 chemical analysis method Methods 0.000 description 1
- 239000003795 chemical substances by application Substances 0.000 description 1
- 230000019771 cognition Effects 0.000 description 1
- 230000004456 color vision Effects 0.000 description 1
- 238000000205 computational method Methods 0.000 description 1
- 238000010924 continuous production Methods 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 230000003247 decreasing effect Effects 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 230000001066 destructive effect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000003292 diminished effect Effects 0.000 description 1
- 230000005284 excitation Effects 0.000 description 1
- 230000001747 exhibiting effect Effects 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000010191 image analysis Methods 0.000 description 1
- 238000002347 injection Methods 0.000 description 1
- 239000007924 injection Substances 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 238000002955 isolation Methods 0.000 description 1
- 238000000691 measurement method Methods 0.000 description 1
- 239000002184 metal Substances 0.000 description 1
- 229910052751 metal Inorganic materials 0.000 description 1
- 238000003801 milling Methods 0.000 description 1
- 238000002156 mixing Methods 0.000 description 1
- 230000001537 neural effect Effects 0.000 description 1
- 238000003921 particle size analysis Methods 0.000 description 1
- 238000004886 process control Methods 0.000 description 1
- 238000010791 quenching Methods 0.000 description 1
- 230000000171 quenching effect Effects 0.000 description 1
- 238000001454 recorded image Methods 0.000 description 1
- 238000009877 rendering Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 239000004460 silage Substances 0.000 description 1
- 238000004611 spectroscopical analysis Methods 0.000 description 1
- 235000019698 starch Nutrition 0.000 description 1
- 239000010959 steel Substances 0.000 description 1
- 238000003860 storage Methods 0.000 description 1
- 239000011031 topaz Substances 0.000 description 1
- 229910052853 topaz Inorganic materials 0.000 description 1
- 238000001429 visible spectrum Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N15/00—Investigating characteristics of particles; Investigating permeability, pore-volume or surface-area of porous materials
- G01N15/02—Investigating particle size or size distribution
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N33/00—Investigating or analysing materials by specific methods not covered by groups G01N1/00 - G01N31/00
- G01N33/24—Earth materials
Definitions
- DIGITAL IMAGE PROCESSING MODELING PARTICULATE PROCESS is applicable to the analysis of all particulate materials such as ores and mixtures, metal surfaces, agricultural grains and biological materials.
- the raw material used to make pellets is iron ore fines.
- Ore processing notably iron ore, such as mine extraction, crushing, milling, transportation, silage, silos removal, moving etc., always generates fines that have no direct application in blast furnaces; on the contrary, they cannot be placed within them.
- deposits whose ores are more friable that is, generate larger amounts of fines, which generate environmental risks and storage costs as well as increased extraction costs to obtain fraction contained within the particle size parameters of ores applicable to blast furnaces.
- the process of producing pellets from ore fines, natural or obtained by grinding has a phase in which by fiotation the silica is eliminated or the content of the silica is greatly reduced.
- the phiotation is performed in its own tanks, being an aqueous process, in which two chemical agents are released in controlled quantities: - starches and amines.
- grayscale images popularly known as black and white
- color images that is, color images.
- a scanned grayscale image corresponds to an array of horizontal and vertical dimensions with pixel values, called pixels.
- a low value pixel corresponds to a dark gray and a high value pixel corresponds to a light gray and thus, as the pixel value increases, one goes to the white representation.
- RGB Redj Green, Blue
- image processing aiming to discriminate objects, recognize shapes, measure sizes and quantities, uses the process of segmentation discrimination that requires the establishment of boundaries between apparent objects in an image and the identification of regions. between borders. For example, to quantify, by volume, a grain type present in a particulate, it is necessary to establish the boundaries between grains, identify the various regions and recognize those regions that correspond to the type of grain sought.
- Borders and identification are operations on pixels, decisions about values of form; Different values, according to a precise decision threshold, delimit borders and similar values identify that they belong to the same region.
- silica grains should be distinguished by their brightness, that is, recognizing lighter and darker grains. Silica grains tend to appear lighter, but the presence of other grains interferes. Grains of hydrated iron ores such as goethite and limonite may appear very light in a grayscale image. Grains of specular hematite, like the very name says, can reflect the incident light like a mirror, even though they are black. This reflection depends on the position and inclination of each grain relative to the incident light and the camera and makes the specular hematite grain to be considered as white in color and as clear as silica and of course to be confused by the processing system.
- Goethite grains may appear reddish, brown, light gray almost white, or yellowish, and are often confused with silica grains. These variations are due to the type of ore being analyzed, ie they depend heavily on which areas of the same mine come from a given sample.
- ores are usually not processed from a single mine area, but mixtures - or blends - of ores from various mine fronts. This means that an identification prepared for samples of one ore type from a given mining front or mix will not serve other types of ores or mixtures.
- sample size that is, the measurement of the sizes of all apparent grains in images by the segmentation and boundary setting process is complex / error-prone and unstable.
- grain size while extremely desirable and useful for controlling grinding and flotation processes, is impractical due to the difficulty of segmenting, establishing the boundaries between grains in images.
- the object of this Patent describes an innovative process in which the quantification of grains in particulate samples is done without the need to segment and / or individually discriminate grains in images absolutely;
- the quantification by images is the result of a model constructed with measurements made on a series of samples by chemical or physical processes, which are used for comparison with the data obtained from the digital images of the samples under analysis.
- samples are selected in sufficient quantity and quality to represent the various situations that may occur in practice.
- samples should be selected from the various mining fronts and "blends" used for the production process, with varying levels of the mineral to be analyzed and also varying particle sizes. This is necessary so that the model that will be responsible for the analysis can capture the universe of possibilities and thus adapt, providing more accurate and stable results.
- the samples are prepared for analysis and quartered, that is, separated into representative aliquots of the same situation and intended for parts for physical or chemical testing and parts for imaging.
- silica chemical tests will be used, eg plasma spectrometry, quantitative mineralogy tests for iron minerals and in sieves for granulometry.
- results are tabulated, forming a data set that will be used as the set of reference variables for modeling.
- the imaging samples are taken to the purpose-built digital imaging apparatus - which is not part of the object of this Patent - where the samples are exposed and the corresponding digital images are recorded. Note that in order to represent each sample well enough picture frames must be recorded. A number of frames around 100 for each sample in general is adequate.
- the image frames thus obtained are analyzed by digital processing and data is extracted from the images and tabulated forming the input data set for the models.
- sample quenching for aliquot separation for physical / chemical testing and imaging is not absolutely necessary. It is described this way because it is common practice in mining, as in the example described.
- the order of physical / chemical testing and imaging operations can be reversed so that the same rates are used for both operations. This inversion will be the form shown in the flow chart for more generality. Also in the case of other particulates, where the tests are not destructive, the same rates may be used for testing and imaging in any order.
- sample set selected for modeling In addition to the sample set selected for modeling, a sample set is also selected for validation, which will be described later.
- the pixels of a digital image can be analyzed in a way that shows the distribution of values in the image. Scan the image by counting the number of pixels that have the same value. The amount of pixels found for each value is divided by the total pixels in the image. A table of correspondence between pixel value and its percentage occurrence is then arranged. This table is called the "Image Histogram". For grayscale images there will be a histogram showing the distribution of values from black to white and for polychromatic images there will be three histograms showing the intensities of R, G and B.
- composition of these three windows must correspond to the color range of the grains of interest, covering all variations of these colors.
- windows do not exactly match the decision thresholds of belonging or not to a particular species, but must match the full range of colors with that the species may appear in the sample set.
- Histograms and, in particular, histogram windows carry quantitative information about the pixel colors present in images without any indication of the positions of these pixels. If an image is fully scrambled and re-analyzed the histograms and window values found will be the same as those found in the previous analysis.
- Histogram windows thus constitute an indirect and global way of representing the colors of objects present in an image without the need to identify each particulate object individually, as is done in the prior art.
- frequency is referred to space and as such is measured in terms of space unit inverses; "per meter” for the relief and "per pixel” for the image, since in the image space is represented in pixels.
- the space in which a repeating pattern occurs is called a loop. For example, if in a picture a light and dark repeating pattern occurs and this pattern has a length of 40 pixels, there is a frequency of 1 cycle per 40 pixels, that is, a frequency of 0.025 cycles / pixel.
- An image containing large objects has lower frequencies at higher intensity and an image containing small objects and many details has higher frequencies at higher intensity.
- An image typically has high and low frequencies to varying degrees depending on the sizes, shapes, and positions of the objects present. Measuring the varying degrees of spatial frequencies constitutes a frequency spectrum. A frequency spectrum will then be an indirect and global way of representing the objects present in an image from their quantities, shapes and positions without the need to identify each object.
- a set of frequencies is selected and the images of the particulate samples are processed to measure the intensities of these frequencies, forming the frequency spectrum.
- the frequency spectrum is the second data set that will form the input variable set for the models. For grayscale images there will be a single spectrum and for color images there will be three spectra corresponding to the images in R, G and B.
- the object of this patent tabulates the values of histogram windows and spatial frequency spectra forming the set of input variables for the models.
- color stimuli and frequency stimuli are processed according to a particular model and respond by identifying the objects in an image.
- This model of interpretation performed automatically by the human brain derives from the individual's learning, which is established about his natural, biological ability to receive stimuli and his neuronal ability to interpret stimuli.
- the model should learn to perform this task by simultaneously placing data from color histogram and frequency spectrum window data as input variables and physical / chemical test data as reference variables for each sample.
- a multivariable linear model operates through parameters, coefficients that link input variables to output variables.
- the modeling phase in this case then consists of calculating these coefficients by specific technique.
- this technique the input and reference data tabulated for all samples are processed together generating this processing a coefficient matrix.
- This matrix then becomes the extract of knowledge, that is, how the model will respond with the contents of granulometry, in the case described, when facing particulate images.
- the digital images of the samples are captured and the data of color histograms and frequency spectra are extracted; On this data, the coefficient matrix is applied, which results in the determination of component contents and their particle size.
- Transformations of image frames prior to calculations of histograms and frequency spectra can also be made which, in some cases, improves the models.
- DIGITAL IMAGE PROCESSING PROCESS OF PARTICULATE ANALYSIS is established and provided to your computer system upon installation for the purpose of analysis of a particular particulate. After the modeling is done, it is necessary to check the adherence of the model, that is, if it is able to respond well to situations other than those used in construction; This phase of deployment is called "model validation”.
- the set of samples selected for validation is treated in the same way as the set used in model construction.
- the process captures the images of these samples, extracts the input values and responds with output values in accordance with the acquired knowledge by applying the coefficient matrix.
- the output values are then compared with the results of the physical / chemical tests, reference values to verify the adherence of the model.
- Figures 1 and 2 correspond, respectively, to the implantation and operation flowcharts of the DIGITAL IMAGE PROCESSING PARTICULATE ANALYSIS PROCESS and Figure 3 to the image capture equipment, object of this patent.
- the implantation consists of eight steps described below and represented in Figure 1.
- the first step is the process of planning (1).
- the process object of this patent is to transport the physical / chemical reality of a particulate to the reality of a computational model of digital images in such a way that the second reality corresponds to the first with fidelity appropriate to its intended purpose.
- the physical / chemical and the image model - lies the effectiveness of the process and the planning must address three fundamental aspects: physical / chemical reality, images and model. It is best to capture the physical / chemical reality, make this reality reflected in digital images and build a model that unites the two realities with the appropriate fidelity to the process.
- Physical / chemical reality may be the content of a component, the particle size, the shape or other attribute of the particles, all considered in this description as the magnitude to be analyzed.
- Planning the capture of physical / chemical reality is the job of knowing the particulate that one wishes to analyze by the process object of this patent.
- the object of this Patent applies to all types of particulates such that the magnitude desired to be analyzed can in some way be perceived in digital images and that there is a physical / chemical method capable of analyzing that same magnitude.
- knowing the particulate and its physical / chemical reality implies the study of the magnitude to be analyzed, the physical / chemical method of analysis and all the forms and values with which the magnitude of interest may occur in practice.
- the second step is to sample preparation (2).
- the plan to capture physical / chemical reality is materialized in the sample set that will be used for modeling.
- the sample set should reflect all forms and values under which the quantities of interest may occur.
- Another set of samples with the same criteria is also prepared, which will be used in the model validation step.
- the third step is the Digital Imaging (3).
- the basic condition that ensures the feasibility of applying the process object of this patent is that the specific quantities of the particles to be analyzed can be perceived in digital images.
- the sharpness of sizes and shapes in digital images is linked to the amount of image pixels that represent an object of interest.
- image magnification that is, the physical size of the image frame and the resolution of the camera.
- the quantity to be analyzed occurs in areas of at least one pixel in the digital image.
- the process object of this patent applies to all types of particulate matter and it is therefore necessary for a given type to study the grain size and size of the digital imaging device to suit the particle size of interest.
- each of the smallest particles of interest appear in the image in at least one pixel.
- a 27mm wide and 20mm high image frame and a 640x480 pixel resolution camera are suitable.
- the colors we assign to objects are a function of the interaction that occurs between the characteristics of the illumination to which they are being subjected with their characteristics. Physically no material has color, but it has the ability to absorb, reflect or depict incident light or emit light when excited by some kind of radiant energy, which may or may not be in the visible spectrum.
- Illumination by visible or infrared light produced by various types of lamps, LEDs or laser is adequate.
- Excitation by non-visible radiation such as ultraviolet or x-rays producing visible light by the fluorescence phenomenon is also adequate.
- the fundamental condition is that digital images are generated in which the quantity of interest to be analyzed reveals itself.
- the samples are taken to the digital imager where the images are taken.
- the images are recorded in a digital file for further processing in the steps of color data extraction (5) and data extraction from spatial frequency spectrum (5).
- the fourth step is to carry out physical tests / chemical (4).
- the object of this patent applies to any type of particulate from which the magnitude to be analyzed may be perceived in digital images in any way and that there is a physical / chemical method capable of analyzing and quantifying the same quantity.
- the quantity to be analyzed may be the content of a component and in this case chemical tests will be performed. If the quantity of interest is the content of a mineral species, quantitative mineralogy tests will be performed. For granuiometry, tests will be made on sieves.
- the magnitude of interest is an analog value - it may be an attribute - and in this case the process object of this Patent will be an analyzer with yes or no answers.
- the process of modeling the quantity of interest, in this case an attribute, with data extracted from images is the same and can be performed as long as there is a method of classifying samples by the attribute in question.
- More than one quantity may be analyzed on the same particulate sample by the process object of this patent. This is the case in the iron ore example where the silica, iron minerals and particle size analysis are analyzed.
- the tests are performed and the results are tabulated and stored temporarily for use in the modeling step (6) as model reference variables.
- the fifth step is the color data extraction (5).
- the recorded images of the particulate samples are, in this step, processed by computer to extract the color data.
- color data is stored for use in modeling step (6) as input variables to models.
- the sixth step is the spectral data extraction spatial frequencies (6).
- the recorded digital images of the particulate samples are, at this stage, processed to extract data from spatial frequency spectra. These data are extracted as a relative intensity of a frequency group, forming a spectrum.
- Frequency values measured in cycles / pixels, and the number of frequencies to make up the spectrum, are chosen to best capture how physical / chemical reality manifested itself in the reality of images in terms of size and shape.
- the frequency set is chosen to best capture the sizes and shapes apparent in the images.
- the "wavelets” technique can be used. (wavelets), the application of specific or other filters. All of these techniques are suitable.
- a set of spatial frequencies that yields good results in many cases is made up of 12 ascending spatial frequencies at diminished fifth intervals, that is, increasing at the square root ratio of two.
- these frequencies correspond to 1 cycle per 80, 56, 40, 28, 20, 14, 10, 7, 5, 4, 3 and 2 pixels.
- the set then starts from the lowest frequency of 1 cycle per 80 pixels, or 0.0125 cycles / pixel, and increases at the square root ratio of two to the highest frequency of 1 cycle per 2 pixels or 0.5 cycles / pixel.
- the intensities of these frequencies are measured by the bandpass filter technique.
- the root ratio of two is approximate to integers, hence the sequence 80, 56, 40, 28, 20, 14, 10, 7, 5, 4, 3 and 2. Note that every two intervals there is twice the frequency, that is, one octave.
- Frequency spectra data extraction is generally performed on each R, G and B image frame. As with colors, you can perform operations on R, G and B frames to obtain transformed frames. with subsequent extraction of frequency data.
- the images are processed and the frequency spectrum data stored for use in the modeling step (6) as input variables to the models.
- the seventh step is the modeling (7).
- the sets, input variables and reference values are processed by a suitable computational technique in order to obtain the model, which constitutes the union between the two realities.
- models can be used such as the multivariable linear model, the nonlinear model, the neural networks and their respective modeling techniques.
- the obtained model consists of the coefficient matrix values, in the case of a multivariable linear model, the definition of operations on the image frames, the definition of the windows on the histograms and the definition of the frequency spectra used. Still for the Model Consolidation (8), it is necessary, before its implementation is completed, to validate the obtained model.
- the obtained model which in principle offers adequate answers to the practical use that will be made of the object of this Patent, must be verified in situations similar to the one according to which it was elaborated, which is called validation of the obtained model.
- samples that have been reserved for this purpose in the sampling step are taken to the digital image capture apparatus for image taking in accordance with the established form of sampling and image capture.
- the values are obtained according to the same definitions of the color histograms and frame transformations and frequency spectrum windows of the obtained model, forming a set of input values for validation.
- the coefficient matrix is applied to each row of the input values, obtaining the corresponding answers from the model.
- the occurrence of a less than adequate hit percentage indicates that the sample set used did not cover all possibilities; In this case, the sample preparation stage should be returned to the generation of new sets that also include the possibilities not previously contemplated.
- the eighth step is consolidation model (8).
- Model Consolidation (8) is performed and the model is stored ("upload") for use in the digital imaging device that will analyze the quantities of interest.
- the consolidated model consists of the coefficient matrix values, in the case of a multivariable linear model, the definition of operations on the image frames, the definition of the windows on the histograms and the definition of the frequency spectra used.
- Model Consolidation (8) in digital memory bit state.
- FIG. 2 is a flowchart showing the use of the process object of this patent, considered and implemented and consolidated the following steps: 1 - use of the first step (9).
- Data from digital processing on sample captures (13) are processed according to the consolidated model and the values of the quantities of interest are inferred.
- results obtained in the previous item may take the form of data for industrial automation system, video information and other applications. typical of data processing.
- this information corresponds to the silica and other mineral contents and the particle size, which are used for flotation process control. 7 - Seventh step of using (15).
- Figure 3 illustrates a schematic diagram of the particulate capture equipment for the example in question where it is used for analysis of silica-containing iron ore particulates.
- a 640x480 pixel color CCD camera with IP connection and iris lens with fixed focus is preferably used.
- It consists of two illuminators with independent intensity settings for each of the elements R, G and B.
- Display table positioned relatively perpendicular to the capture camera.
- a table of sufficient size to capture a 27mm wide and 20mm frame reading is suitable.
Landscapes
- Chemical & Material Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Pathology (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- General Health & Medical Sciences (AREA)
- Engineering & Computer Science (AREA)
- Immunology (AREA)
- Physics & Mathematics (AREA)
- Dispersion Chemistry (AREA)
- Environmental & Geological Engineering (AREA)
- General Life Sciences & Earth Sciences (AREA)
- Geology (AREA)
- Remote Sensing (AREA)
- Food Science & Technology (AREA)
- Medicinal Chemistry (AREA)
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
Description
Claims
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
PCT/BR2014/000389 WO2016061647A1 (pt) | 2014-10-21 | 2014-10-21 | Processo de análise de particulados via modelagem por processamento de imagens digitais |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
PCT/BR2014/000389 WO2016061647A1 (pt) | 2014-10-21 | 2014-10-21 | Processo de análise de particulados via modelagem por processamento de imagens digitais |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2016061647A1 true WO2016061647A1 (pt) | 2016-04-28 |
Family
ID=55759965
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/BR2014/000389 WO2016061647A1 (pt) | 2014-10-21 | 2014-10-21 | Processo de análise de particulados via modelagem por processamento de imagens digitais |
Country Status (1)
Country | Link |
---|---|
WO (1) | WO2016061647A1 (pt) |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
GB2302736A (en) * | 1995-06-29 | 1997-01-29 | Ibm | Estimating grain size in geological samples |
US20080192235A1 (en) * | 2002-07-26 | 2008-08-14 | Olympus Corporation | Image processing system |
-
2014
- 2014-10-21 WO PCT/BR2014/000389 patent/WO2016061647A1/pt active Application Filing
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
GB2302736A (en) * | 1995-06-29 | 1997-01-29 | Ibm | Estimating grain size in geological samples |
US20080192235A1 (en) * | 2002-07-26 | 2008-08-14 | Olympus Corporation | Image processing system |
Non-Patent Citations (4)
Title |
---|
BIANCONI, F. ET AL.: "Theoretical and experimental comparison of different approaches for colour texture classification", Retrieved from the Internet <URL:http://antfdez.webs.uvigo.es/publications/2011_JEI.pdf> * |
CUSANO, C. ET AL.: "Intensity and color descriptors for texture classification, Proc. of SPIE- IS &T Electronic Imaging", SPIE, vol. 8661, 20 May 2015 (2015-05-20), Retrieved from the Internet <URL:http://proceedings.spiedigitallibrary.org/pdfaccess.ashx?ResourceID=5383099&PDFSource=24> * |
LU , L. ET AL.: "Terrain surface classification with a control mode update rule using a 2D laser stripe-based structured light sensor", ROBOTICS AND AUTONOMOUS SYSTEMS, vol. 59, 2011, pages 954 - 965, XP028293187, Retrieved from the Internet <URL:http://ac.elscdn.com/S0921889011001205/1-s2.0-S0921889011001205-main.pdf'?-tid=439ee7aa-ffl4-1le4-a8de-00000aabO06c&acdnat=1432142502_a01e8fd1b3260668ae7b9e5e087fne06> [retrieved on 20150520], doi:10.1016/j.robot.2011.06.015 * |
OLIVEIRA, E. F. ET AL.: "Granulometric analysis based on the energy of Wavelet Transform coefficients", REM: R. ESC. MINAS, OURO PRETO, vol. 63, no. 2, April 2010 (2010-04-01), pages 347 - 354, Retrieved from the Internet <URL:http://www.scielo.br/pdf/rem/v63n2/rem63n2a20.pdf> * |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US11861889B2 (en) | Analysis device | |
US11893731B2 (en) | Group sparsity model for image unmixing | |
Arboleda | Comparing performances of data mining algorithms for classification of green coffee beans | |
Macfarlane et al. | Automated estimation of foliage cover in forest understorey from digital nadir images | |
Izadi et al. | A new intelligent method for minerals segmentation in thin sections based on a novel incremental color clustering | |
CN105427275B (zh) | 大田环境麦穗计数方法及装置 | |
JP2006526783A (ja) | 色分解および組織試料内の対象となる領域の検出 | |
TWI642126B (zh) | 半導體晶圓檢測系統及其方法 | |
CN110163101B (zh) | 中药材种子区别及等级快速判别方法 | |
CN105842173A (zh) | 一种高光谱材质鉴别方法 | |
Font et al. | An image processing method for in-line nectarine variety verification based on the comparison of skin feature histogram vectors | |
CN107563427A (zh) | 用于油画的著作权鉴定的方法以及相应的使用 | |
Jacq et al. | Sedimentary structure discrimination with hyperspectral imaging in sediment cores | |
Gurubasava et al. | Analysis of agricultural soil pH using digital image processing | |
Mohan et al. | An intelligent recognition system for identification of wood species | |
Portalés et al. | An image-based system to preliminary assess the quality of grape harvest batches on arrival at the winery | |
Ramakrishnan et al. | Visual dictionaries as intermediate features in the human brain | |
TWI517055B (zh) | Image foreground object screening method | |
WO2016061647A1 (pt) | Processo de análise de particulados via modelagem por processamento de imagens digitais | |
Han et al. | Research on grading detection of the wheat seeds | |
Afor et al. | Color Detection System for Diamond Sorting Using Machine Learning | |
DE112021001620T5 (de) | Teilchenanalysesystem und teilchenanalyseverfahren | |
Gourav | Rice Grain Quality Determination Using OTSU's Thresholding Method | |
Sigit et al. | Performance analysis of color matching technique for teeth classification based on color histogram | |
Sarkate et al. | Domain specific knowledge based machine learning for flower classification using soft computing |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 14904320 Country of ref document: EP Kind code of ref document: A1 |
|
DPE1 | Request for preliminary examination filed after expiration of 19th month from priority date (pct application filed from 20040101) | ||
NENP | Non-entry into the national phase |
Ref country code: DE |
|
REG | Reference to national code |
Ref country code: BR Ref legal event code: B01A Ref document number: 112017008319 Country of ref document: BR |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 14904320 Country of ref document: EP Kind code of ref document: A1 |