# US 2024/0122229 A1 · Method for evaluating cigarette burning quality indicators

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

- Number: US 2024/0122229 A1
- Document ID: US20240122229A1
- Status: Published application
- Kind: A1
- Application number: 18/504,248
- Filed: 2023-11-08
- Published: 2024-04-18
- Earliest claimed: 2022-07-07
- Assignees: China Tobacco Yunnan Industrial Co., Ltd
- Applicants: China Tobacco Yunnan Industrial Co., Ltd.
- Inventors: Han ZHENG; Jianbo ZHAN; Hao WANG; Zhenhua YU; Xu WANG; Jiao XIE; Ying ZHANG; Tao WANG; Tingting YU; Baoshan YUE
- Web page: https://patentgrep.com/patent/US20240122229A1

## CPC

- A24C 5/3406: {Controlling cigarette combustion}
- G01N 21/84: Systems specially adapted for particular applications
- G01J 3/0248: {using a sighting port, e.g. camera or human eye}
- G06T 2207/30108: Industrial image inspection
- A24C: MACHINES FOR MAKING CIGARS OR CIGARETTES
- G01N: INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES (measuring or testing apparatus or processes other than immunoassay, involving enzymes or microorganisms C12M, C12Q)
- G01J: MEASUREMENT OF INTENSITY, VELOCITY, SPECTRAL CONTENT, POLARISATION, PHASE OR PULSE CHARACTERISTICS OF INFRARED, VISIBLE OR ULTRAVIOLET LIGHT; COLORIMETRY; RADIATION PYROMETRY (light sources F21, H01J, H01K, H05B; investigating properties of materials by optical means G01N)
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL

## Abstract

A method for comprehensively analyzing and/or evaluating cigarette burning quality index is disclosed. The steps include collecting cigarette burning quality index data, filtering the cigarette burning quality index data, standardizing cigarette data, and measuring cigarette burning quality. The importance of the cigarette burning quality indicators can also be evaluated. The method for comprehensively analyzing and/or evaluating cigarette burning quality indicators can reflect general laws more accurately by maintaining the sample distribution through a singularity detection method, and analyzing correlations of each index with cigarette performance from multiple perspectives, to fuse them into a comprehensive measurement value. The importance ranking and weight of indicators can be obtained more completely and stably.

## Claims

### Claim 1

A method for comprehensively analyzing and/or evaluating cigarette burning quality indicators, comprising:
while simulating smoking using a robot under predefined environmental simulation conditions, capturing real-time status of combusting cigarettes using a plurality of cameras at different angles for the cigarette burning quality indicators, where the cigarette burning quality indicators include a crack rate, an ash column deviation, an ash column gray level, an ash column length, an ash column area, an ash shrinkage rate, a fly ash value, an ash retention rate, a charcoal line width, a charcoal line uniformity and a burning speed;
establishing a subsample isolated tree, wherein the subsample isolated tree includes a subsample number and a tree height satisfying:
ht=ceiling(log2(ψ))
where ht is the smallest integer value greater than or equal to log2(ψ), which is approximately an average path length of the subsample isolated tree:
c
⁡
(
m
)
=
2
⁢
H
⁡
(
m
-
1
)
-
2
⁢
(
m
-
1
)
m
;
where m is a number of samples, c(m) is the average path length or an average height of the subsample isolated tree, and H(m−1) is a harmonic number;
constructing an isolated forest containing a number of the subsample isolated trees ct;
passing the samples through each the subsample isolated tree in the isolated forest to obtain an anomaly score of each sample, where the anomaly score of an individual sample x is calculated according to:
s
⁡
(
x
,
m
)
=
2
-
E
⁡
(
h
⁡
(
x
)
)
c
⁡
(
m
)
where h(x) is a path length of the sample x, E(h(x)) is an expectation of h(x), and an anomaly score range of the sample is (0,1);
removing ones of the samples having an anomaly score above a predetermined threshold, and denoting data after removing the ones of the samples as Xnorm=(x1norm, x2norm, . . . , x11norm);
dividing the cigarette burning quality indicators are divided into positive indicators and negative indicators;
processing the positive indicators as follows:
x
i
=
x
inorm
-
min
⁡
(
x
inorm
)
max
⁡
(
x
inorm
)
-
min
⁡
(
x
inorm
)
,
i
=
1
,
2
,
…
,
11
;
processing the negative indicators as follows
x
i
=
max
⁡
(
x
inorm
)
-
x
inorm
max
⁡
(
x
inorm
)
-
min
⁡
(
x
inorm
)
,
i
=
1
,
2
,
…
,
11
;
where X=(x1, x2, . . . , x11) is standard cigarette burning quality data, and max(xinorm) and min(xinorm) are respectively maximum and minimum values of each indicator in the data after removing the ones of the samples;
combining a grayscale correlation, a Euclidean distance, a city block distance and a cosine similarity analysis of the data after removing the ones of the samples, wherein a grayscale correlation coefficient X between each of the cigarette burning quality indices and the corresponding data after removing the ones of the samples is:
ζ
⁢
i
⁡
(
k
)
=
min
s
min
t
❘
"\[LeftBracketingBar]"
x
0
(
t
)
-
x
s
(
t
)
❘
"\[RightBracketingBar]"
+
ρ
max
s
max
t
❘
"\[LeftBracketingBar]"
x
0
(
t
)
-
x
s
(
t
)
❘
"\[RightBracketingBar]"
❘
"\[LeftBracketingBar]"
x
0
(
k
)
-
x
i
(
k
)
❘
"\[RightBracketingBar]"
+
ρ
max
s
max
t
❘
"\[LeftBracketingBar]"
x
0
(
t
)
-
x
s
(
t
)
❘
"\[RightBracketingBar]"
;
where ζi(k) is the grayscale correlation coefficient, ρ is a resolution coefficient,
min
s
min
t
❘
"\[LeftBracketingBar]"
x
0
(
t
)
-
x
s
(
t
)
| is a two-level minimum difference,
min
s
min
t
❘
"\[LeftBracketingBar]"
x
0
(
t
)
-
x
s
(
t
)
| is a two-level maximum difference, and obtaining a grayscale correlation value of the data after removing the ones of the samples and a corresponding cigarette sample performance reference column according to:
d
r
=
1
n
⁢
∑
k
=
1
n
ζ
i
(
k
)
;
the Euclidean distance between the cigarette burning quality index and the corresponding cigarette sample performance reference column is:
d
e
=
∑
i
=
1
n
(
x
i
-
x
0
)
2
;
the city block distance between the cigarette burning quality index and the corresponding cigarette sample performance reference column is:
d
c
=
∑
i
=
1
n
❘
"\[LeftBracketingBar]"
x
i
-
x
0
❘
"\[RightBracketingBar]"
;
and
the cosine similarity between the cigarette burning quality index and the corresponding cigarette sample performance reference column is:
d
cos
=
1
-
X
·
x
0

X

×

x
0

=
1
-
∑
i
=
1
n
x
i
×
x
0
∑
i
=
1
n
x
i
2
×
∑
i
=
1
n
x
0
2
;
and
combining the grayscale correlation, the Euclidean distance, the city block distance and the cosine similarity analysis according to:
d
=
d
r
-
min
⁡
(
d
r
)
(
max
⁡
(
d
r
)
-
min
⁡
(
d
r
)
)
+
∑
i
=
1
2
1
d
i
-
min
⁡
(
1
d
i
)
2
⁢
(
max
⁡
(
1
d
i
)
-
min
⁡
(
1
d
i
)
)
+
1
d
cos
-
min
⁡
(
1
d
cos
)
(
max
⁡
(
1
d
cos
)
-
min
⁡
(
1
d
cos
)
)
;
where d is a comprehensive measure, dr is a grayscale correlation degree; when i=1,2, d1 and d2 represent the Euclidean distance and city block distance respectively, and dcos represents the cosine similarity.

### Claim 2 (depends on 1)

The method according to claim 1, wherein in Step S21, the number of sub-samples ψ is 256.

### Claim 3 (depends on 1)

The method according to claim 1, wherein in step S22, ct=100.
