Apparatus and process for optimizing radiation detection counting times using machine learning

Patent metadata

Field	Value
Publication number	US-11249199-B2
Application number	US-201916978156-A
Country	US
Kind code	B2
Filing date	Mar 4, 2019
Priority date	Mar 16, 2018
Publication date	Feb 15, 2022
Grant date	Feb 15, 2022

Abstract

A method is provided to reduce the counting times in radiation detection systems using machine learning, wherein the method comprises: receiving an output data from a detector which is to detect a target material from a target body; analyzing the output data; identifying a material of interest from the analyzed output data; and controlling a source of the target material to prevent the source from harming the target body. An apparatus is also provided which comprises: a detector to detect radiation and to provide an output data in real-time; and a processor coupled to the detector, wherein the processor is to: receive the output data; analyze the output data; identify a material of interest from the analyzed output data; and control a source of the target material.

First claim

What is claimed is:

1. An apparatus to reduce counting times in radiation detection systems, the apparatus comprising:
a detector to detect a target material in real-time and to provide output data; and
a processor coupled to the detector, wherein the processor is to:
receive the output data;
analyze the output data in real-time, wherein the processor is to count a material in the output data in real-time to analyze the output data in real-time, and wherein the processor is to collect time-stamped counting data off a multi-channel analyzer of a counting system to analyze the output data;
identify a material of interest from the analyzed output data; and
control a source of the target material, wherein the target material is classified according to two or more machine-learning training models, and wherein the two or more machine-learning training models are applied in parallel, sequential, or a combination of them.

2. The apparatus of claim 1 , wherein the processor is to generate energy histograms from quantified available energy readings.

3. The apparatus of claim 2 , wherein the processor is to classify the target material based on the energy histograms.

4. The apparatus of claim 3 , wherein the processor is to identify the target material, and to send an instruction to the source to control the source of the target material.

5. The apparatus of claim 2 , wherein the two or more machine-learning training models include a Bayesian statistical model and a Markov model.

6. The apparatus of claim 1 , wherein the two or more machine-learning training models are of same type but trained on different data.

7. The apparatus of claim 5 , wherein the two or more machine-learning training models are of differing types but trained on same data.

8. The apparatus of claim 5 , wherein the two or more machine-learning training models are of differing types and are trained on different data.

9. The apparatus of claim 1 , wherein the target material is a radiation material.

10. The apparatus of claim 1 , wherein to control the source of the target material comprises to stop the source from radiating or to reduce radiation from the source.

11. A machine-readable storage media having machine-readable instruction that, when executed, cause one or more machines to perform a method, the method comprising:
receiving an output data from a detector which is to detect a target material from a target body;
analyzing the output data in real-time, wherein the one or more machines is to count a material in the output data in real-time to analyze the output data in real-time, and wherein the one or more machines is to collect time-stamped counting data off a multi-channel analyzer of a counting system to analyze the output data;
identifying a material of interest from the analyzed output data; and
controlling a source of the target material to prevent the source from harming the target body, wherein the target material is classified according to two or more machine-learning training models, and wherein the two or more machine-learning training models are applied in parallel, sequential, or a combination of them.

12. The machine-readable storage media of claim 11 , wherein analyzing the output data comprises:
quantifying available energy readings;
generating energy histograms from the quantified available energy readings; and
classifying the target material based on the energy histograms.

13. The machine-readable storage media of claim 12 , wherein the two or more machine-learning training models include a Bayesian statistical model and a Markov model.

14. The machine-readable storage media of claim 11 , wherein controlling the source of the target material comprises stopping the source from radiating or reducing radiation from the source.

15. The machine-readable storage media of claim 11 , wherein the target body is an organic body, and wherein the target material is a radiation material.

16. A system comprising:
a radiation source to radiate a target material;
a detector to detect the target material in real-time and to provide an output data;
a processor coupled to the detector, wherein the processor is to:
receive the output data;
analyze the output data, wherein the processor is to count a material in the output data to analyze the output data, and wherein the processor is to collect time-stamped counting data off a multi-channel analyzer of a counting system to analyze the output data;
identify a material of interest from the analyzed output data; and
control a source of the target material, wherein the target material is classified according to two or more machine-learning training models, and wherein the two or more machine-learning training models are applied in parallel, sequential, or a combination of them; and

a display to display the analyzed output data.

17. The system of claim 16 wherein the processor is to:
quantify available energy readings;
generate energy histograms from the quantified available energy readings;
classify the target material based on the energy histograms and/or the two or more machine-learning training models;
identify the target material; and
send an instruction to the source to control the source of the target material.

Assignees

• Univ Oregon State

Inventors

• Frieder Ophir
• Reese Steven Richard
• Curtis Jessica Ryan

Classifications

• G01T1/17

  Circuit arrangements not adapted to a particular type of detector {(pulse-selection circuits H03K, G01R)} · CPC title

• G01T1/18 Primary

  with counting-tube arrangements, e.g. with Geiger counters (tubes H01J47/08; {with alarm provision G01T7/125}) · CPC title

• G01T1/36

  Measuring spectral distribution of X-rays or of nuclear radiation {spectrometry (pulse selection circuits per se H03K; investigation of materials by radiation diffraction G01N23/20; spectrometer tubes H01J49/00)} · CPC title

• G01V5/22 Primary

  Active interrogation, i.e. by irradiating objects or goods using external radiation sources, e.g. using gamma rays or cosmic rays · CPC title