# US 12,599,934 B2 · Multiple stage sorting

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 12,599,934 B2
- Document ID: US12599934B2
- Status: Granted
- Kind: B2
- Application number: 18/731,134
- Filed: 2024-05-31
- Published: 2024-09-19
- Granted: 2026-04-14
- Earliest claimed: 2015-07-16
- Assignees: Sortera Technologies, Inc.
- Applicants: Sortera Technologies Inc.
- Inventors: Nalin Kumar; Manuel Gerardo Garcia, Jr.
- Web page: https://patentgrep.com/patent/US12599934B2

## CPC

- B07C 5/3422: {using video scanning devices, e.g. TV-cameras}
- B07C 5/34: Sorting according to other particular properties {(material testing per se G01N; quality control G07C3/14)}
- B07C 5/342: according to optical properties, e.g. colour {(according to radiation transmittivity B07C5/3416)}
- B07C 5/04: Sorting according to size {(sorting devices as a component of the automatic supply and removal of workpieces B23Q7/12)}
- B07C 2501/0054: Sorting of waste or refuse
- B07C: POSTAL SORTING; SORTING INDIVIDUAL ARTICLES, OR BULK MATERIAL FIT TO BE SORTED PIECE-MEAL, e.g. BY PICKING (specially adapted for a specific purpose covered by another class, see the relevant place, e.g. A43D33/06, B23Q7/12)

## Abstract

A material sorting system sorts materials utilizing multiple stages of classification and sorting, including a vision system that implements an artificial intelligence system in order to identify or classify each of the materials, and an x-ray fluorescence (“XRF”) system or Laser Induced Breakdown Spectroscopy to perform a subsequent classification and sorting of the remaining materials.

## Claims

### Claim 1

A system for handling a first mixture of materials comprising a plurality of different classes of materials, the system comprising:
an image sensor configured to capture visually observed characteristics of each of the first mixture of materials; and
a data processing system comprising a machine learning system implementing a neural network configured with a previously generated set of neural network parameters to classify a first plurality of materials of the first mixture as belonging to a first class of materials based solely on the captured visually observed characteristics, wherein the previously generated set of neural network parameters are uniquely associated with the first class of materials, wherein the plurality of materials of the first mixture classified as belonging to the first class of materials possess a chemical composition that is different from the materials within the first mixture not classified as belonging to the first class of materials.

### Claim 2 (depends on 1)

The system as recited in claim 1, wherein the previously generated set of neural network parameters uniquely associated with the first class of materials were generated from captured visually observed characteristics of one or more samples of the first class of materials.

### Claim 3 (depends on 1)

The system as recited in claim 1, wherein the first class of materials is cast aluminum alloys, the system further comprising:
a first sorter configured to sort the classified first plurality of materials of the first mixture from the first mixture as a function of the classifying of the first plurality of materials of the first mixture, wherein the sorting by the first sorter of the classified first plurality of materials of the first mixture from the first mixture produces a second mixture of materials that comprises the first mixture minus the classified first plurality of materials of the first mixture;
a Laser Induced Breakdown Spectroscopy (“LIBS”) system configured to classify a second plurality of materials of the second mixture as belonging to a second class of materials; and
a second sorter configured to sort the classified second plurality of materials of the second mixture from the second mixture as a function of the classifying of the second plurality of materials of the second mixture by the LIBS system, wherein the second mixture of materials comprises wrought aluminum material pieces containing a plurality of different wrought aluminum alloys, and wherein the LIBS system is configured to classify certain ones of the second mixture as belonging to a first wrought aluminum alloy, wherein the second sorter sorts the classified certain ones from the second mixture as a function of the classifying of certain ones of the second mixture, wherein the sorting by the second sorter of the classified certain ones from the second mixture produces a third mixture of materials that comprises the second mixture minus the certain ones from the second mixture, wherein the third mixture comprises materials belonging to a second wrought aluminum alloy different from the first wrought aluminum alloy.

### Claim 4 (depends on 1)

The system as recited in claim 1, wherein the first class of materials is cast aluminum alloys, the system further comprising:
a first sorter configured to sort the classified first plurality of materials of the first mixture from the first mixture as a function of the classifying of the first plurality of materials of the first mixture;
an x-ray fluorescence (“XRF”) system configured to classify a second plurality of materials of the classified first plurality of materials as belonging to a second class of materials as a function of spectral data produced by the XRF system; and
a second sorter configured to sort the classified second plurality of materials from the classified first plurality of materials as a function of the classifying of the second plurality of materials by the XRF system.

### Claim 5 (depends on 1)

The system as recited in claim 1, wherein the previously generated set of neural network parameters were produced in a training stage in which an artificial intelligence system implementing a neural network processed visual images of a control set of materials representing the first class of materials.

### Claim 6

A method for handling a first heterogeneous mixture of separable materials comprising a plurality of different types of materials, the method comprising:
capturing visually observed characteristics of each material piece of the first heterogeneous mixture of materials with a camera;
assigning, with an artificial intelligence system implementing a neural network configured with a previously generated set of neural network parameters, a first classification to certain ones of the first heterogeneous mixture of materials as belonging to a first type of materials based solely on the captured characteristics of each material piece of the first heterogeneous mixture of materials, wherein the previously generated set of neural network parameters are uniquely associated with the first type of materials; and
sorting the certain ones of the first heterogeneous mixture of materials from the first heterogeneous mixture as a function of the first classification, wherein the sorting produces a second heterogeneous mixture of materials that comprises the first heterogeneous mixture of materials minus the sorted certain ones of the first heterogeneous mixture of materials.

### Claim 7 (depends on 6)

The method as recited in claim 6, wherein the previously generated set of neural network parameters were produced from a previously generated classification of a control sample of the first type of materials.

### Claim 8 (depends on 6)

The method as recited in claim 6, further comprising:
assigning with a LIBS system a second classification to certain ones of the second heterogeneous mixture of materials as belonging to a second type of materials; and
sorting the certain ones of the second heterogeneous mixture of materials from the second heterogeneous mixture as a function of the second classification.

### Claim 9 (depends on 8)

The method as recited in claim 8, wherein the first classification of materials is cast aluminum alloys, wherein the second heterogeneous mixture of materials comprises wrought aluminum material pieces containing a plurality of different wrought aluminum alloys, and wherein the LIBS system is configured to classify certain ones of the second heterogeneous mixture as belonging to a first wrought aluminum alloy, wherein a sorter sorts the classified certain ones of the second heterogeneous mixture as a function of the classifying of certain ones of the second heterogeneous mixture.

### Claim 10 (depends on 9)

The method as recited in claim 9, wherein the sorting by the sorter of the classified certain ones of the second heterogeneous mixture produces a third mixture of materials that comprises the second heterogeneous mixture minus the certain ones of the second heterogeneous mixture, wherein the third mixture comprises materials belonging to a second wrought aluminum alloy different from the first wrought aluminum alloy.

### Claim 11 (depends on 8)

The method as recited in claim 8, wherein the second heterogeneous mixture of materials comprises metal cast alloys.

### Claim 12 (depends on 6)

The method as recited in claim 6, further comprising:
assigning with an XRF system a second classification to certain ones of the second heterogeneous mixture of materials as belonging to a second type of materials as a function of spectral data produced by the XRF system; and
sorting the certain ones of the second heterogeneous mixture of materials from the second heterogeneous mixture as a function of the second classification.

### Claim 13 (depends on 12)

The method as recited in claim 12, wherein the first classification of materials is cast aluminum alloys.

### Claim 14 (depends on 12)

The method as recited in claim 12, wherein the second heterogeneous mixture of materials comprises metal cast alloys.

### Claim 15 (depends on 6)

The method as recited in claim 6, wherein the previously generated set of neural network parameters were produced in a training stage in which an artificial intelligence system implementing a neural network processed visual images of a control set of materials representing the first class of materials.