BE1029449B1 - MULTI-LAYER FILM MONITORING - Google Patents

Abstract

Multi-layered film monitoring by merged analyzes includes acquiring a spectroscopic dataset of a sample object and performing both AM analysis of the dataset and also chemometric analysis of the sample object. set of images. The duplicate analyzes are then merged into a single output specifying a composition of each of a multiplicity of constituent components of the sample object, for example a specification of a presence of one or more polymers such as acrylonitrile-butadiene- styrene (ABS), polystyrene (PS), polyethylene (PE), polycarbonate (PC), polypropylene (PP) or polyamide (PA).

Description

MULTILAYER FILM MONITORING

BACKGROUND OF THE INVENTION

Field of invention

The present invention relates to the technical field of the recycling of thermoplastic composites and more particularly to the dissolution of multilayer thermoplastics.

Description of prior art

Plastic recycling is the process of recovering plastic scrap or waste and reprocessing the material into useful products. The nature of plastic recycling depends on the type of plastic itself. Unlike a thermoset plastic, a thermoplastic, or heat malleable plastic, is a plastic polymer material that becomes bendable or moldable at a certain high temperature and solidifies upon cooling. The recycling of thermoplastics includes many benefits such as the supply of raw materials for the manufacturing industry, a reduced environmental threat for humans since they are non-biodegradable, incineration and landfill problems reduced to a minimum, consumption lower energy for sustenance, and it serves as a source of income and employment opportunities.

There are several methods for recycling thermoplastics, including primary recycling, mechanical recycling, and chemical recycling. In chemical recycling, waste plastics serve as raw materials and are converted into monomer or other products by decomposition and depolymerization of the starting material through the use of thermal energy or a catalyst. In the case of chemical — recycling of a thermoplastic, chemical dissolution decomposition and depolymerization results in the decomposition of waste polymeric thermoplastics into lower molecular weight species for reuse. Dissolution can be achieved by immersing the thermoplastic in a solvent such as benzene, chlorobenzene, trichlorethylene, toluene and xylene.

In order to correctly define an optimal process for the recovery of a polymer constituent component of interest from thermoplastic waste, the identity of the constituent component must be known a priori. Traditionally, recognition of the polymeric constituent component is achieved using hyperspectral imaging followed by chemometric analysis of the imagery produced by the hyperspectral imaging. However, it is widely understood that while chemometrics works well in the case of a single-layer plastic, chemometrics does not work well when it comes to a multi-layered plastic film.

BRIEF SUMMARY OF THE INVENTION

Embodiments of the present invention address technical deficiencies in the art of monitoring multilayer films using chemometric analysis. To this end, the embodiments of the present invention provide a novel and non-obvious method of multi-layer film monitoring by merged analyses. The embodiments of the present invention also make it possible to obtain a new and non-obvious computing device suitable for carrying out the previous method. Finally, the embodiments of the present invention allow to obtain a new and non-obvious data processing system incorporating the previous device in order to realize the previous method.

In one embodiment of the invention, a method for multilayer film monitoring by fused scans includes acquiring a spectroscopic data set of a sample object, such as by means of hyperspectral imaging (HSI) or laser-induced plasma spectroscopy (LIBS). Then both AM analysis of the dataset and also chemometric analysis of the dataset can be performed. Therefore, the duplicate analyzes can be merged into a single output specifying a composition of each of a multiplicity of constituent components of the sample object. Specifically, AM analysis of the image set includes subjecting a data structure representative of imagery in the data set to a rule-based inference decision tree. In this regard, the decision tree can be trained with a set of sample images each having been annotated with a ground truth indication of a corresponding material composition of different constituent components, for example a polymer such as acrylonitrile-butadiene-styrene (ABS), polystyrene (PS), polyethylene (PE), polycarbonate (PC), polypropylene (PP) or polyamide (PA). Therefore, the decision tree is able to return for each image in the set a probability of the image exhibiting a specific set of the different components among the constituent components each of the corresponding components of the corresponding material compositions.

In one aspect of the embodiment, the merging includes receiving a first indication for each image in the data set from the chemometric analysis of a first set of constituent components each of a determined first composition of the corresponding material composition and also receiving a second indication for each image in the data set from the AM analysis of a second set of constituent components each of a first determined composition of the material composition corresponding. Next, the first and second indications are statistically combined for each of the constituent components in a union of the first set and the second set to produce a merged set of the constituent components of a statistically merged determination of the corresponding material composition of each of the constituent components overall merged.

In another aspect of the embodiment, the merging includes receiving a first indication for each image in the data set from the chemometric analysis of a first set of constituent components each of a first determined composition of the corresponding material composition and also receiving a second indication for each image in the data set from the AM analysis of a second set of constituent components each of a determined first composition of the composition of corresponding material. Then, one of the first set of constituent components and the second set of constituent components can be selected as a merged set of the constituent components according to a highest probability of characterization produced by the chemometric analysis and the AM analysis, respectively. .

In another embodiment of the invention, a data processing system is suitable for multilayer film monitoring by fused analyses. The system includes a spectroscopic imaging sensor, such as an HSI camera or a LIBS sensor, a host computing platform that has one or more computers, each having memory, and one or more processing units including one or more processing cores and a communication channel established between the sensor and one of the computers on which the spectroscopic data captured by the sensor is received in the memory of the host computing platform. The system also includes a multi-layer film monitoring module by merged analyses. The module includes computer program instructions validated while being executed in the memory of at least one of the processing units of the host computing platform to acquire a set of spectroscopic data of a sample object from the sensor from of the communication channel, to perform a dual AM and chemometric analysis of the data set and to merge the dual analyzes into a single output specifying a composition of each of a plurality of constituent components of the sample object.

In yet another embodiment of the invention, a computing device includes a non-transitory computer-readable storage medium having program instructions stored therein. The instructions are executable by at least one processing core of a processing unit to cause the processing unit to perform a multilayer film monitoring method by merged analyses. The method includes acquiring a spectroscopic dataset from a sample object and performing dual AM and chemometric analysis of the dataset. Finally, the dual scans are merged into a single output specifying a composition of each of a multiplicity of 5 constituent components of the sample object.

In this way, the technical deficiencies of the singular use of chemometrics in the characterization of constituent components of multilayer plastic waste are overcome due to the fusion of a chemometric analysis with an AM analysis of a multilayer plastic film. Also, since the spectroscopic sensor is a LIBS sensor, materials of all colors can be characterized with greater accuracy compared to an HSI camera which does not allow characterization of black-colored materials.

Additional aspects of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Aspects of the invention will be realized and achieved by means of the elements and combinations particularly pointed out in the appended claims. It is understood that the preceding general description and the following detailed description are both given by way of example and explanation only and do not limit the invention, as claimed.

BRIEF DESCRIPTION OF THE SEVERAL VIEWS OF THE DRAWINGS

The accompanying drawings, which are incorporated into and form a part of this description, illustrate embodiments of the invention and together with the description, serve to explain the principles of the invention. The embodiments illustrated herein are preferred herein, it being understood, however, that the invention is not limited to the specific arrangements and instruments shown, in which:

Figure 1 is a pictorial illustration reflecting different aspects of a multi-layered film monitoring process by fused scans;

Figure 2 is a block diagram representing a data processing system suitable for carrying out one of the aspects of the method of Figure 1; And,

Figure 3 is a flowchart illustrating one aspect of the process for

Figure 1.

DETAILED DESCRIPTION OF THE INVENTION

Embodiments of the invention provide multi-layered film monitoring by fused scans. In accordance with one embodiment of the invention, a multilayer film of thermoplastic waste may be subjected to spectroscopic imaging to produce a spectroscopic data set of the waste. The data set is then subjected to a chemometric analysis in order to classify according to a degree of probability one or more constituent components by corresponding material composition. Simultaneously, the image set is also subjected to AM analysis by submitting the data set to a trained classifier to classify the image set in a degree of probability according to one or more constituent components per material composition . Classifications from both chemometric analysis and AM analysis are then merged to more accurately classify the constituent components of the images despite the multi-layered nature of the waste.

By way of illustration of one aspect of the embodiment, Figure 1 pictorially shows a method of multi-layer film monitoring by merged scans.

As shown in Figure 1, a spectroscopic imaging sensor 110, for example an HIS camera or a LIBS sensor, scans a multilayer plastic film 100 including a prospective combination of one or more polymeric materials, including ABS,

PS, PE, PC, PP or PA. Scanning by the imaging sensor 110 produces a spectroscopic data set 130 of different images 120 corresponding to different frequencies. Next, the spectroscopic data set 130 is subjected to chemometric analysis 140 so as to produce a characterization 160 of the constituent components of the multilayer film 100.

Simultaneously, the spectroscopic data set 130 is subjected to AM analysis 150 which includes a rule-based inference decision tree trained to correlate different image features with an annotated material among the polymeric materials. As such, the AM analysis 150 produces a characterization 170 of the constituent components of the multilayer film 100 and includes an accuracy probability for the characterization 170. A merging process 180 then merges the characterizations 170 to produce a merged characterization 190 of the constituent components of the multilayer film 100 to be displayed to an end user.

For example, the fusion process 180 for each of the images 120 in the data set 130 can compare the characterization 160 from the chemometric analysis 140 to the characterization 170 from the AM analysis 150 and combine the two as a union of the constituent components predicted by the characterizations 160, 170. Alternatively, the merging process 180 for each of the images 120 in the data set 130 may compare the characterization 160 of the chemometric analysis 140 to the characterization 170 of the AM analysis 150 and select one of the characterizations 160, 170 having, attributed thereto, a highest probability. As yet another variation, the merging process 180 for each of the images 120 in the data set 130 may compare the characterization 160 from the chemometric analysis 140 to the characterization 170 from the AM analysis 150 and weight each predicted constituent component of each of the characterizations according to an assigned probability, filtering out any predicted constituent component of a probability lower than a threshold value.

The aspects of the method described in connection with Figure 1 can be implemented inside a data processing system. In a further illustration, Figure 2 schematically shows a data processing system suitable for performing multilayer film monitoring by merged analyses. In the data processing system illustrated in Figure 1, a host computing platform 200 is provided. Host computing platform 200 includes one or more computers 210, each having memory 220 and one or more processing units 230. Host computing platform computers 210 (only a single computer shown for simplicity of illustration) may be co-located within and in communication with each other over a local area network, or over a data communication bus, or the computers may be disposed remote from and in communication with each other with each other through a network interface 260 over a data communications network. It is important to note that a spectroscopic imaging sensor 270 such as a LIBS sensor or an HIS camera is communicatively coupled to the host computing platform 200 over a data communication interface 240, such as a serial cable or a parallel data bus.

Notably, a computing device 250 including a non-transitory computer-readable storage medium may be included with the data processing system 200 and accessible by the processing units 230 of one or more of the computers 210. The computing device 250 stores thereon or encloses therein a program module 300 communicatively coupled to logic for performing chemometric analysis 280A, and also a rule-based inference decision tree 280B trained to classify an image based on a training data set of a plurality of images each annotated with a ground truth indication of a classification of constituent components of a particular material. Notably, the program module 300 includes computer program instructions which, when executed by one or more of the processing units 230, perform a programmatically executable method for multi-layer film monitoring by fused scans.

Specifically, the program instructions during execution receive from the spectroscopic imaging sensor 270 on the data communication interface 240, a set of data 290 captured for a multilayer plastic film.

The program instructions then submit the data set 290 to each of the chemometric analysis 280A and the decision tree 280B in order to receive from each a classification of the constituent components of the multilayer plastic film visible in the assembly. of images 290. The program instructions then merge the classification from each of the chemometric analysis 280A and the convolutional neural network 280B to produce a merged characterization of the constituent components of the multi-layered plastic film to be displayed in the host computing platform 200 .

In this regard, the program instructions may merge the characterizations according to a highest probability of the presence of each of the constituent components, or based on a combination of the determined constituent components having an associated probability exceeding a minimum threshold, or the program instructions may augment the Characterization of the chemometric analysis 280A with the constituent components indicated by the convolutional neural network 280B as being absent in the characterization produced by the chemometric analysis 280A. In this way, the relative inaccuracies of the independent use of the 280A chemometric analysis can be resolved with an analysis

AM such as that produced by convolutional neural network 280B.

In a further illustration of an exemplary operation of the module, Figure 3 is a flowchart illustrating one of the method aspects of the

Figure 1. Beginning in block 310, a spectroscopic dataset of a multi-layered plastic film is received from a spectroscopic imaging sensor and in block 320, the dataset imagery is processed by chemometric analysis 320 in parallel with an AM analysis in block 330. In block 340, the chemometric analysis produces a characterization of the multilayer plastic film in terms of the material composition of the various constituent components of the multilayer plastic film. Similarly, in block 350, the AM analysis produces a characterization of the multilayer plastic film. In block 360, a union of the characterized constituent components of the two sets is made to create a merged characterization. Finally, in block 370, the merged characterization is displayed to the end user.

It is important to note that the preceding flow chart and block diagram referred to herein illustrate the architecture, functionality, and operation of possible implementations of computing systems, methods, and devices according to various embodiments of the present invention. In this regard, each block in the flowchart or block diagrams may represent a module, segment or part of instructions, which includes one or more executable instructions to implement the specified logical function or functions. In some alternate implementations, the functions noted in the block may occur in an order different from the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially simultaneously, or the blocks may sometimes be executed in reverse order, depending on the functionality involved. It is also noted that each block in the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, may be implemented by specialized hardware-based systems that perform specified functions or acts or perform combinations of specialized hardware and computer instructions.

More specifically, the present invention may be embodied in a programmatically executable method. Similarly, the present invention may be embodied in a computing device on which programmatic instructions are stored and from which the programmatic instructions are enabled to be loaded into a memory of a data processing system and executed from the latter in order to carry out the above programmatically executable method. Further, the present invention may be embodied in a data processing system adapted to load the programmatic instructions from a computing device and then to execute the programmatic instructions to perform the foregoing programmatically executable method.

For this purpose, the computing device is a non-transitory computer-readable storage medium or mediums containing therein or storing thereon computer-readable program instructions. These instructions, when executed from memory by one or more processing units of a data processing system, cause the processing units to perform various programmatic processes which are examples of various aspects of the executable process of programmatically. In this regard, processing units each include an instruction executing device such as a central processing unit or “CPU” of a computer. One or more computers may be included within the data processing system. It should be noted that while the CPU may be a single-core CPU, it is understood that multiple CPU cores may operate within the CPU and in either case the instructions are directly loaded. from memory in one or more of the cores of one or more of the CPUs for execution.

Besides loading instructions directly from memory for execution by one or more cores of a CPU or multiple CPUs, the computer-readable program instructions described herein may alternatively be retrieved from a computer communication network into the memory of a computer of the data processing system for execution therein. Similarly, only part of the program instructions can be retrieved into memory from the computer communication network, while other parts can be loaded from persistent storage of the computer. Furthermore, only part of the program instructions can be executed by one or more processing cores of one or more CPUs of one of the computers of the data processing system, while other parts can be executed cooperative within a computer different from the data processing system which is either co-located with the computer or positioned remotely from the computer on the computer communication network with the results of the calculation by the two computers shared between these.

The corresponding structures, materials, acts and equivalents of all means or step plus function elements in the claims below are intended to include any structure, material or act for performing the function in combination with other claimed elements such as specifically claimed. The description of the present invention has been presented for purposes of illustration and description but is not intended to be exhaustive or limited to the invention in the form disclosed. Many modifications and variations will be apparent to those skilled in the art without departing from the scope and spirit of the invention. The embodiment has been selected and described in order to better explain the principles of the invention and practical application, and to enable others skilled in the art to understand the invention for various embodiments having various modifications as adapted to the particular intended use.

Having thus described the invention of the present application in detail and with reference to the embodiments thereof, it will be apparent that modifications and variations are possible without departing from the scope of the invention defined in the claims. appended as follows:

Claims (15)

CLAIMS We claim: 1. Multi-layer film monitoring method by merged analyzes including: acquisition of a spectroscopic data set of a sample object; performing dual machine learning (ML) and chemometric analysis of the dataset; and, merging the dual scans into a single output specifying a composition of each of a multiplicity of constituent components of the sample object. 2. A method according to claim 1, wherein the spectroscopic data set is a data set produced by Laser Induced Plasma Atomic Emission Spectroscopy (LIBS). The method of claim 1, wherein the AM analysis of the dataset includes subjecting a data structure representative of an imagery of the dataset to a rule-based inferential decision tree trained with a set of sample images each annotated with a ground truth indication of a corresponding material composition of different ones of the constituent components, the decision tree returning for each image the imagery in the set of data a probability that said image exhibits a specific set of the different ones of the constituent components, each of the corresponding components of the corresponding material composition. A method according to claim 3, wherein merging comprises: receiving a first indication for each of said images in the data set from chemometric analysis of a first set of constituent components each of a first determined composition of the corresponding material composition; receiving a second indication for each of said images in the data set from the AM analysis of a second set of constituent components each of a first determined composition of the corresponding material composition; and, statistically combining the first and second indications for each of the constituent components in a union of the first set and the second set to produce a merged set of the constituent components of a statistically merged determination of the corresponding material composition of each constituent components in the merged assembly. A method according to claim 3, wherein merging comprises: receiving a first indication for each of said images in the data set from chemometric analysis of a first set of constituent components each of a first determined composition of the corresponding material composition; receiving a second indication for each of said images in the data set from the AM analysis of a second set of constituent components each of a first determined composition of the corresponding material composition; and, selecting one of the first set of constituent components and the second set of constituent components as the merged set of constituent components according to a highest probability of characterization produced by the chemometric analysis and the AM analysis, respectively . 6. A data processing system adapted for multilayer film monitoring by fused analyses, the system comprising: a spectroscopic imaging sensor; a host computing platform comprising one or more computers, each comprising memory and one or more processing units including one or more processing cores; a communication channel established between the sensor and one of the computers on which image data captured by the camera is received in the memory of the host computing platform; and, a fused analysis multi-layer film monitoring module comprising computer program instructions activated during execution in the memory of at least one of the processing units of the host computer platform to carry out: the acquisition of a set of spectroscopic data of a sample object from the sensor via the communication channel; performing dual machine learning (ML) and chemometric analysis of the dataset; and, merging the dual scans into a single output specifying a composition of each of a multiplicity of constituent components of the sample object. 7. The system of claim 6, wherein the sensor is a Laser Induced Plasma Atomic Emission Spectroscopy (LIBS) sensor. The system of claim 6, wherein the AM analysis of the dataset includes subjecting a data structure representative of an imagery of the dataset to a rule-based inferential decision tree trained with a set of sample images each annotated with a ground truth indication of a corresponding material composition of different ones of the constituent components, the decision tree returning for each image the imagery in the set of data a probability that said image exhibits a specific set of the different ones of the constituent components, each of the corresponding components of the corresponding material composition. The system of claim 8, wherein merging comprises: receiving a first indication for each of said images in the data set from chemometric analysis of a first set of constituent components each of a first determined composition of the corresponding material composition; receiving a second indication for each of said images in the data set from AM analysis of a second set of constituent components each of a first determined composition of the corresponding material composition; and, statistically combining the first and second indications for each of the constituent components in a union of the first set and the second set to produce a merged set of the constituent components of a statistically merged determination of the corresponding material composition of each constituent components in the merged assembly. The system of claim 8, wherein merging comprises: receiving a first indication for each of said images in the data set from chemometric analysis of a first set of constituent components each of a first determined composition of the corresponding material composition; receiving a second indication for each of said images in the data set from the AM analysis of a second set of constituent components each of a first determined composition of the corresponding material composition; and, selecting one of the first set of constituent components and the second set of constituent components as the merged set of constituent components according to a highest probability of characterization produced by the chemometric analysis and the AM analysis, respectively . 11. Computing device comprising a computer-readable non-transitory storage medium in which program instructions are stored, the instructions being executable by at least one processing core of a processing unit to cause the processing unit to perform a a method of multilayer film monitoring by fused scans, the method including: acquiring a spectroscopic data set of a sample object; performing dual machine learning (ML) and chemometric analysis of the dataset; and, merging the dual scans into a single output specifying a composition of each of a multiplicity of constituent components of the sample object. 12. … A computing device according to claim 11, wherein the spectroscopic data set is a data set produced by Laser Induced Plasma Atomic Emission Spectroscopy (LIBS). 13. A computing device according to claim 11, wherein the AM analysis of the dataset comprises subjecting a data structure representative of an imagery of the dataset to an inferential decision tree based on rules trained with a set of sample images each annotated with a ground truth indication of a corresponding material composition of different ones of the constituent components, the decision tree returning for each image the imagery in the data set a probability that said image exhibits a specific set of the different ones of the constituent components each of the corresponding components of the corresponding material composition. 14. A computing device according to claim 13, wherein merging comprises: receiving a first indication for each of said images in the data set from chemometric analysis of a first set of constituent components each of a first determined composition of the corresponding material composition; receiving a second indication for each of said images in the data set from the AM analysis of a second set of constituent components each of a first determined composition of the corresponding material composition; and, statistically combining the first and second indications for each of the constituent components in a union of the first set and the second set to produce a merged set of the constituent components of a statistically merged determination of the corresponding material composition of each constituent components in the merged assembly. 15. A computing device according to claim 13, wherein merging comprises: receiving a first indication for each of said images in the data set from chemometric analysis of a first set of constituent components each of a first determined composition of the corresponding material composition; receiving a second indication for each of said images in the data set from the AM analysis of a second set of constituent components each of a first determined composition of the corresponding material composition; and, selecting one of the first set of constituent components and the second set of constituent components as the merged set of constituent components according to a highest probability of characterization produced by the chemometric analysis and the AM analysis, respectively .