CN104034847B - A kind of accurate smell fingerprint detection method based on fast discrete frequency-domain analysis theory - Google Patents
A kind of accurate smell fingerprint detection method based on fast discrete frequency-domain analysis theory Download PDFInfo
- Publication number
- CN104034847B CN104034847B CN201310742360.4A CN201310742360A CN104034847B CN 104034847 B CN104034847 B CN 104034847B CN 201310742360 A CN201310742360 A CN 201310742360A CN 104034847 B CN104034847 B CN 104034847B
- Authority
- CN
- China
- Prior art keywords
- signal
- discrete frequency
- frequency domain
- domain
- detection method
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Fee Related
Links
Landscapes
- Investigating Or Analyzing Materials By The Use Of Fluid Adsorption Or Reactions (AREA)
Abstract
The invention discloses a kind of accurate smell fingerprint detection method based on fast discrete frequency-domain analysis theory, comprise gather the sample process of gas data, truncation, signal period continuation process, discrete frequency domain convert four steps.The advantage that the present invention has and good effect are: the accurate smell fingerprint detection method based on the conversion of fast discrete frequency domain of the present invention, the disturbing molecule that observation data is concentrated can effectively be proposed, and greatly can improve operation efficiency, and by rejecting disturbing molecule to the frequency-domain and time-domain analysis of disturbing molecule, improve the accuracy of detection.Therefore, the present invention not only enhances the real-time of systems axiol-ogy, and testing result is more reliable, greatly improves the accuracy of detection.The present invention is applicable to fixing or hand-hold electric nasus, can greatly improves the technical indicator such as detection speed and accuracy.
Description
Technical field
The invention belongs to monitored gas environment technical field, particularly relate to a kind of accurate smell fingerprint detection method based on fast discrete frequency-domain analysis theory.
Background technology
In recent years, China's haze weather is more and more serious, and the number of times of annual haze also constantly increases, and has had a strong impact on life and the health of people.And the vehicle exhaust etc. in the waste gas causing the main cause of haze to be exactly industrial gaseous waste to get rid of as chemical plant, steel-making and oil refining etc. and life; for due to these other the feature such as heterogeneity and unevenness; make the monitoring of people to harmful gas face a very large difficult problem, and the real-time of monitoring and the information such as accuracy successfully manage people and control crisis and all have very important left and right.
Occurred a kind of new smell fingerprint detection method in recent years in Electronic Nose market, this system mainly utilizes the smell sensors array of device interior and gas molecular data processing and identification device to detect gas finger-print.In the entire system, the effect of Electronic Nose is equivalent to the olfactory organ of people, other molecular signals received are converted after treatment to the data point in gas data space, then, foundation is to harmful gas scientific research data setting dependent thresholds, the gas data accepted is classified, the class of harmful gas is fed back to system front end, trigger relative alarm.
Gas detecting system can arrange the chemical composition data of the gas molecule that sensor array receives, to the effective sorting technique of gas data Information Pull after arrangement, harmful gas is classified with safe gas, then, sorted result is calculated to concentration or the ratio of harmful gas in current scene, according to this feature identification harmful gas.Owing to there is a lot of interference in air, therefore, to the key that specific dusty gas molecule detects, be just how to detect quickly and accurately from these gas molecules and pollute molecule, this just proposes very high technical requirement to the Processing Algorithm of smell sensor.
For existing traditional harmful gas checkout equipment, due to the general not high reason of process data volume macrooperation efficiency, make the real-time poor effect detected in actual monitoring.In addition, due to the limitation of traditional Processing Algorithm, make the detection accuracy of harmful gas poor, so, existing monitoring equipment due to its real-time and accuracy poor, also cannot meet actual monitoring needs far away.
Summary of the invention
The present invention is in order to solve exist in above-mentioned existing detection method consuming time long, accurately can not detect the technical barriers such as harmful gas, provide a kind of accurate smell fingerprint detection method based on fast discrete frequency-domain analysis theory, the method is applied in Electronic Nose, the disturbing molecule that observation data is concentrated can effectively be proposed, and greatly operation efficiency be can improve, the current real-time of smell fingerprint detection and the requirement of accuracy effectively improve.
The present invention for addressing this problem taked technical scheme is:
Accurate smell fingerprint detection method based on fast discrete frequency-domain analysis theory of the present invention, main treatment scheme comprise gather the sample process of gas data, truncation, signal period continuation process, discrete frequency domain convert four steps.Concrete methods of realizing is as follows:
The first step, the sample process of gas data
The signal array gathered by smell sensor array is carried out sampling processing.Its sample frequency is set to
, sampling interval is
, wherein,
.
Second step, signal cutout process
After sampling processing is carried out to the signal gathered, then carry out signal cutout process, suppose that truncated signal is:
Then by signal cutout to interval
after signal representation be:
3rd step, signal period continuation process
By the olfactory signal data translation through over-sampling and truncation
formed cyclical signal, by with
carry out convolution to realize this process.Therefore can be expressed as through the mathematical expression of periodic extension signal:
Wherein: time-domain signal
4th step: discrete frequency domain converts
If the odor data collected
for length is the finite sequence of N, so can define
n point discrete frequency domain be transformed to:
And its inverse transformation can be expressed as:
Wherein:
.
The advantage that the present invention has and good effect are:
Accurate smell fingerprint detection method based on the conversion of fast discrete frequency domain of the present invention, the disturbing molecule that observation data is concentrated can effectively be proposed, and greatly can improve operation efficiency, and by rejecting disturbing molecule to the frequency-domain and time-domain analysis of disturbing molecule, improve the accuracy of detection.The present invention is applicable to fixing or hand-hold electric nasus, can greatly improves the technical indicator such as detection speed and accuracy.
Accompanying drawing explanation
Fig. 1 is the decomposable process flow graph based on the conversion of fast discrete frequency domain of the present invention.
Embodiment
Referring to drawings and Examples, the accurate smell fingerprint detection method based on the conversion of fast discrete frequency domain of the present invention is described in detail.
Accurate smell fingerprint detection method based on the conversion of fast discrete frequency domain of the present invention, embodiment is as follows:
The first step, the sample process of gas data
The signal array gathered by smell sensor array is carried out sampling processing.Its sample frequency is set to
, sampling interval is
, wherein,
.
Second step, signal cutout process
After sampling processing is carried out to the signal gathered, then carry out signal cutout process, suppose that truncated signal is:
Then by signal cutout to interval
after signal representation be:
3rd step, signal period continuation process
By the olfactory signal data translation through over-sampling and truncation
formed cyclical signal, by with
carry out convolution to realize this process.Therefore can be expressed as through the mathematical expression of periodic extension signal:
Wherein: time-domain signal
4th step: discrete frequency domain converts
If the odor data collected
for length is the finite sequence of N, so can define
n point discrete frequency domain be transformed to:
And its inverse transformation can be expressed as:
Wherein:
.
For fast discrete frequency domain mapping algorithm, important technological difficulties are implementation procedures of this algorithm, and main in the present invention what adopt is fast discrete frequency-domain transform method, and so the circular of this method is:
It is as follows that frequency domain converts concrete computation process:
For discrete series
if,
, and satisfy condition
Wherein:
Then:
Therefore can obtain:
Due to:
Therefore:
Can obtain based on the above results:
Then, the form of the operation result of above formula with signal flow diagram can be represented, calculating process as shown in Figure 1.The beneficial outcomes applying this decomposition algorithm generally can improve operation efficiency, and more comprehensively can analyze collection signal, improves the accuracy to image data classification, and then improve the detection accuracy of system.
Computation process operand divides
For
's
the conversion of some discrete frequency domain, needs after decomposition
individual computing level and
the conversion of some discrete frequency domain, each computing level needs answering to take advantage of and be added with computing with twice once, and each
the discrete frequency domain conversion of point needs altogether
secondary take advantage of again and
secondaryly be added with computing, after decomposing, total computing comprises altogether
secondary take advantage of again and
secondaryly to be added with.
So the operand after decomposing decreases about half, then, can also the sequence after decomposition be decomposed further.By each
the subsequence of point is being two by Parity-decomposition
the subsequence of point obtains:
Then
In like manner can obtain:
According to above decomposable process, if
time, the operation times that this algorithm needs altogether is:
Take advantage of again:
Be added with:
And the calculated amount of directly carrying out discrete frequency domain conversion is:
Take advantage of again:
Be added with:
.
By above to computing quantitative analysis, can find out when carrying out frequency domain conversion to a large amount of data, adopt rapid computations more can save time and space than Direct Transform, the beneficial outcomes be applied in physical device is: the operand that generally can reduce system, greatly improve operation efficiency, strengthen the real-time of equipment.In addition, by more comprehensively analyzing the signal of frequency domain after conversion, be more convenient for the classification of collection signal, more effectively can reject the impact of undesired signal on testing result, ensure that the accuracy of systems axiol-ogy.
Claims (2)
1., based on an accurate smell fingerprint detection method for fast discrete frequency-domain analysis theory, it is characterized in that, the method comprise gather the sample process of gas data, truncation, signal period continuation process, discrete frequency domain convert four steps; Concrete methods of realizing is as follows:
The first step, the sample process of gas data
The signal array gathered by smell sensor array is carried out sampling processing; Its sample frequency is set to
, sampling interval is
, wherein,
;
Second step, signal cutout process
After sampling processing is carried out to the signal gathered, then carry out signal cutout process, if truncated signal is:
Then by signal cutout to interval
after signal representation be:
3rd step, signal period continuation process
By the olfactory signal data translation through over-sampling and truncation
formed cyclical signal, by with
carry out convolution to realize this process; Therefore be expressed as through the mathematical expression of periodic extension signal:
Wherein: time-domain signal
4th step: discrete frequency domain converts
If the odor data collected
for length is the finite sequence of N,
n point discrete frequency domain be transformed to:
And its inverse transformation can be expressed as:
Wherein:
.
2. the accurate smell fingerprint detection method based on fast discrete frequency-domain analysis theory according to claim 1, is characterized in that: the detailed process of discrete frequency domain conversion is as follows:
For discrete series
if,
, and satisfy condition
Wherein:
Then:
Therefore can obtain:
Due to:
Therefore:
Can obtain based on the above results:
For
's
the conversion of some discrete frequency domain, needs after decomposition
individual computing level and
the conversion of some discrete frequency domain, each computing level needs answering to take advantage of and be added with computing with twice once, and each
the discrete frequency domain conversion of point needs altogether
secondary take advantage of again and
secondaryly be added with computing, after decomposing, total computing comprises altogether
secondary take advantage of again and
be added with;
Then, the sequence after decomposition is decomposed further; By each
the subsequence of point is being two by Parity-decomposition
the subsequence of point obtains:
Then
In like manner can obtain:
。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310742360.4A CN104034847B (en) | 2013-12-30 | 2013-12-30 | A kind of accurate smell fingerprint detection method based on fast discrete frequency-domain analysis theory |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310742360.4A CN104034847B (en) | 2013-12-30 | 2013-12-30 | A kind of accurate smell fingerprint detection method based on fast discrete frequency-domain analysis theory |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104034847A CN104034847A (en) | 2014-09-10 |
| CN104034847B true CN104034847B (en) | 2016-03-30 |
Family
ID=51465695
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310742360.4A Expired - Fee Related CN104034847B (en) | 2013-12-30 | 2013-12-30 | A kind of accurate smell fingerprint detection method based on fast discrete frequency-domain analysis theory |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104034847B (en) |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101840001A (en) * | 2010-02-10 | 2010-09-22 | 中国科学院地质与地球物理研究所 | Acquiring method and device of geological structure three-dimensional imaging data |
| CN102594373A (en) * | 2011-01-07 | 2012-07-18 | 北京中科国技信息***有限公司 | Method for generating low-complexity SSB (Single Side Band) signals of RFID (Radio Frequency Identification Device) system |
| CN103293553A (en) * | 2013-04-17 | 2013-09-11 | 中国海洋石油总公司 | Continuation and correction method for boundary element of earthquake data collected through upper cables and lower cables in complex seabed |
| CN103440871A (en) * | 2013-08-21 | 2013-12-11 | 大连理工大学 | Method for suppressing transient noise in voice |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9075159B2 (en) * | 2011-06-08 | 2015-07-07 | Chevron U.S.A., Inc. | System and method for seismic data inversion |
-
2013
- 2013-12-30 CN CN201310742360.4A patent/CN104034847B/en not_active Expired - Fee Related
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101840001A (en) * | 2010-02-10 | 2010-09-22 | 中国科学院地质与地球物理研究所 | Acquiring method and device of geological structure three-dimensional imaging data |
| CN102594373A (en) * | 2011-01-07 | 2012-07-18 | 北京中科国技信息***有限公司 | Method for generating low-complexity SSB (Single Side Band) signals of RFID (Radio Frequency Identification Device) system |
| CN103293553A (en) * | 2013-04-17 | 2013-09-11 | 中国海洋石油总公司 | Continuation and correction method for boundary element of earthquake data collected through upper cables and lower cables in complex seabed |
| CN103440871A (en) * | 2013-08-21 | 2013-12-11 | 大连理工大学 | Method for suppressing transient noise in voice |
Non-Patent Citations (3)
| Title |
|---|
| 基于时域和频域信号综合分析的金属探伤研究;章学铜 等;《机电工程》;20130331;第30卷(第3期);全文 * |
| 潘 楠 等.基于频域盲解卷积的机械设备状态监测与故障诊断.《振 动 与 冲 击》.2012,第 ３１卷(第 １２期),全文. * |
| 用频域方法实时计算运动中的脉搏数;吴 剑 等;《清华大学学报 (自然科学版 )》;20021231;第42卷(第3期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104034847A (en) | 2014-09-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110319982B (en) | Buried gas pipeline leakage judgment method based on machine learning | |
| CN103901162B (en) | Gas detecting system and method in a kind of portable vehicle | |
| CN103631681B (en) | A kind of method of online reparation abnormal data of wind power plant | |
| CN201780678U (en) | Online monitoring system for various factors in water quality | |
| CN109633094B (en) | Online odor concentration monitoring method | |
| CN109767054A (en) | Efficiency cloud appraisal procedure and edge efficiency gateway based on deep neural network algorithm | |
| CN103412557A (en) | Industrial fault detection and diagnostic method suitable for nonlinear process on-line monitoring | |
| CN205080114U (en) | Portable pollution sources monitoring system based on satellite positioning | |
| CN116933044A (en) | Intelligent processing method and system for power supply data | |
| CN116415723A (en) | Wind speed prediction method and device for railway disaster prevention system | |
| CN116187861A (en) | Isotope-based water quality traceability monitoring method and related device | |
| CN104034847B (en) | A kind of accurate smell fingerprint detection method based on fast discrete frequency-domain analysis theory | |
| CN1896742A (en) | Space-pollution realtime monitoring indicator | |
| CN102890718A (en) | Electronic nose data mining method based on supervised explicit manifold learning algorithm | |
| CN114492146B (en) | Bolt group loosening positioning and quantitative analysis method and system based on transfer learning | |
| CN115598164A (en) | Machine learning integrated soil heavy metal concentration detection method and system | |
| CN103325065A (en) | Decision-making method for sampling for detection of quality safety of agricultural products | |
| CN203249888U (en) | Device for detecting mixed gas of hydrogen sulfide and carbon monoxide | |
| CN111582734A (en) | Ocean pollution comparative analysis and risk assessment intelligent method based on python crawler system and SVM | |
| CN106709467A (en) | Real-time spike potential detection and classification method suitable for hardware realization | |
| Wang et al. | Signal disaggregation via sparse coding with featured discriminative dictionary | |
| CN205091232U (en) | Haze pollution sources monitoring devices | |
| CN109614954B (en) | Qualitative classification method for hyperspectral migration of heavy metal pollution in soil | |
| CN103973494A (en) | Alarm filtering method of cloud data center | |
| Makar et al. | Source Attribution of VOCs in the Canadian Oil Sands using Hierarchical Clustering |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20160330 Termination date: 20171230 |