Attached substance determination method, attached substance determination device, and non-transitory computer-readable storage medium storing attached substance determination program

Patent metadata

Field	Value
Publication number	US-11262315-B2
Application number	US-202017130488-A
Country	US
Kind code	B2
Filing date	Dec 22, 2020
Priority date	Dec 23, 2019
Publication date	Mar 1, 2022
Grant date	Mar 1, 2022

Abstract

An attached substance determination method causes a computer to determine whether or not an attached substance is attached to an inspection target object, in which the computer includes at least one processor, and the at least one processor is configured to (a) acquire, as learning data, a spectroscopic image obtained by imaging a first type sample having the attached substance attached to a base with a spectroscopic camera and a spectroscopic image obtained by imaging a second type sample having no attached substance attached to the base with the spectroscopic camera, in which spectroscopic images of a plurality of kinds of the first type samples having different kinds of the bases and different kinds of the attached substances and spectroscopic images of a plurality of kinds of the second type samples having different kinds of the bases are acquired as the learning data, (b) generate, based on the learning data, a determination model with a spectroscopic image of the inspection target object as an input and a determination result relating to presence or absence of the attached substance as an output, (c) acquire the spectroscopic image of the inspection target object, and (d) input the spectroscopic image of the inspection target object to the determination model and determine the presence or absence of the attached substance based on the determination result output from the determination model.

First claim

What is claimed is:

1. An attached substance determination method that causes a computer to determine whether or not an attached substance is attached to an inspection target object, wherein
the computer includes at least one processor, and
the at least one processor is configured to
(a) acquire, as learning data, a spectroscopic image obtained by imaging a first type sample having the attached substance attached to a base with a spectroscopic camera and a spectroscopic image obtained by imaging a second type sample having no attached substance attached to the base with the spectroscopic camera, in which spectroscopic images of a plurality of kinds of the first type samples having different kinds of the bases and different kinds of the attached substances and spectroscopic images of a plurality of kinds of the second type samples having different kinds of the bases are acquired as the learning data,
(b) generate, based on the learning data, a determination model with a spectroscopic image of the inspection target object as an input and a determination result relating to presence or absence of the attached substance as an output,
(c) acquire the spectroscopic image of the inspection target object, and
(d) input the spectroscopic image of the inspection target object to the determination model and determine the presence or absence of the attached substance based on the determination result output from the determination model.

2. The attached substance determination method according to claim 1 , wherein
the at least one processor is configured to, in (a),
acquire, as the learning data, spectroscopic images of a plurality of kinds of the first type samples having different colors of the bases and spectroscopic images of a plurality of kinds of the second type samples having different colors of the bases.

3. The attached substance determination method according to claim 1 , wherein
the at least one processor is configured to, in (a),
acquire, as the learning data, spectroscopic images of a plurality of kinds of the first type samples having different colors of the attached substances.

4. An attached substance determination device comprising:
at least one processor, wherein
the at least one processor is configured to
(a) acquire, as learning data, a spectroscopic image obtained by imaging a first type sample having an attached substance attached to a base with a spectroscopic camera and a spectroscopic image obtained by imaging a second type sample having no attached substance attached to the base with the spectroscopic camera,
(b) generate, based on the learning data, a determination model with a spectroscopic image of an inspection target object as an input and a determination result relating to presence or absence of the attached substance as an output,
(c) acquire the spectroscopic image of the inspection target object, and
(d) input the spectroscopic image of the inspection target object to the determination model and determine the presence or absence of the attached substance based on the determination result output from the determination model, and
the at least one processor is configured to further
acquire, as the learning data, spectroscopic images of a plurality of kinds of the first type samples in which kinds of the base and the attached substance of the first type sample are changed and spectroscopic images of a plurality of kinds of the second type samples in which the kind of the base is changed.

5. A non-transitory computer-readable storage medium storing instructions causing at least one processor to execute:
(a) acquiring, as learning data, a spectroscopic image obtained by imaging a first type sample having the attached substance attached to a base with a spectroscopic camera and a spectroscopic image obtained by imaging a second type sample having no attached substance attached to the base with the spectroscopic camera, in which spectroscopic images of a plurality of kinds of the first type samples having different kinds of the bases and different kinds of the attached substances and spectroscopic images of a plurality of kinds of the second type samples having different kinds of the bases are acquired as the learning data;
(b) generating, based on the learning data, a determination model with a spectroscopic image of the inspection target object as an input and a determination result relating to presence or absence of the attached substance as an output;
(c) acquiring the spectroscopic image of the inspection target object; and
(d) inputting the spectroscopic image of the inspection target object to the determination model and determining the presence or absence of the attached substance based on the determination result output from the determination model.

Assignees

• Seiko Epson Corp

Inventors

• Kanemoto Kei
• Kanai Masashi

Classifications

• G06N3/0464

  Convolutional networks [CNN, ConvNet] · CPC title

• G06N3/09

  Supervised learning · CPC title

• G06N3/08

  Learning methods · CPC title

• G01N2021/8883

  involving the calculation of gauges, generating models · CPC title

• G01J3/2823

  Imaging spectrometer · CPC title