CN105158178A - Rapid modeling method for detecting sugar content of navel orange based on spectral peak area in high spectral transmission technology - Google Patents
Rapid modeling method for detecting sugar content of navel orange based on spectral peak area in high spectral transmission technology Download PDFInfo
- Publication number
- CN105158178A CN105158178A CN201510644358.2A CN201510644358A CN105158178A CN 105158178 A CN105158178 A CN 105158178A CN 201510644358 A CN201510644358 A CN 201510644358A CN 105158178 A CN105158178 A CN 105158178A
- Authority
- CN
- China
- Prior art keywords
- navel orange
- sample
- spectrum
- hyperion
- pol
- 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.)
- Granted
Links
Landscapes
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
The invention discloses a rapid modeling method for detecting the sugar content of a navel orange based on the spectral peak area in a high spectral transmission technology, and relates to the technical field of nondestructive detection of the fruit internal quality. The rapid modeling method comprises the following steps: (1) obtaining semi-transmissive hyperspectral patterns of a navel orange sample; (2) detecting the sugar content of the navel orange sample according to a chemical method; (3) selecting the average hyperspectral pattern of the navel orange sample; (4) calculating the spectral peak area of the average hyperspectral pattern; (5) establishing a sugar content predictive model for the navel orange sample, and utilizing the model for quality detection. The rapid modeling method has the advantages that the hyperspectral patterns of the navel orange sample are acquired in a semi-transmissive manner based on the high spectral transmission technology, so that the internal quality information of the navel orange sample can be effectively obtained, and the detection level and the detection efficiency for the fruit internal quality can be improved; the rapid modeling method is high in modeling efficiency, accuracy and model calculation speed, and can be utilized for rapidly detecting the sugar content and other indexes of the fruit internal quality, and then evaluating the fruit internal quality.
Description
Technical field
The present invention relates to fruit quality technical field of nondestructive testing, particularly relate to a kind of navel orange pol based on EO-1 hyperion through-transmission technique spectrum peak area and detect rapid modeling method.
Background technology
China is Production of fruit big country, occupies first place in the world in the long-term position of output, accounts for 19% of world's fruit total production.Development of Fruit Industry is rapid, and often present the trend risen every year, fruit industry has developed into the mainstay industry of the many local peasant's extra earnings of China, has promote greatly and development function agricultural and economy.Although Shi Guo industry big producing country of China, export volume is still very little on the whole, not fruit powerful country of commerce.Citrus fruit is the large fruit of the first in the world, and the current orange yield of China occupies the 3rd (accounting for 10.8%) between countries in the world, and front two is Brazil (accounting for 23.7%) and U.S.'s (accounting for 15.9%) respectively; Up to the present, China's fruit exit amount does not also exceed 4% of fruit year total production.Meanwhile, along with the raising of internal people's level of consumption, the consumption status of consumer is by the requirement requirement of product volume being evolved to confrontation, when buying fruit, people not only only focus on the external sort of fruit, for the inside quality of fruit, especially as the requirement also corresponding raising of the inside qualities such as pol, mouthfeel and nutritional labeling.In order to promote that fruit is marketed, improving product increment, carrying out reprocessing and reprocessing after namely commercial treatment gathers to fruit to fruit, realize fruit quality, promote the market competitiveness, expand export, improve application and the indispensable means of operator's economic benefit; Wherein how carrying out Fast nondestructive evaluation to fruit inside quality is the important step of carrying out a series of fruit commercial treatment.
Because machine vision technique and spectral technique have fast, can't harm and reliable advantage, be used widely in nondestructive measuring method of the farm product at present.High light spectrum image-forming technology is the integration technology of a kind of image and spectrum, can obtain space and the spectral information of detected object simultaneously.View data and spectroscopic data combine, and can be good at, in conjunction with pointed the obtaining interior of articles physical arrangement and chemical composition information of surface, having good application prospect.But concerning navel orange class pachydermia fruit, more difficult acquisition internal information, high spectrum image information data amount is comparatively large, adds difficulty and the time of data processing simultaneously.
Summary of the invention
In order to solve navel orange skin depth, internal information difficulty obtains, hyperspectral image data amount is large, be difficult to accurately detect the difficult problems such as its pol fast in based on the quick online detection process of hyper-spectral image technique, the invention provides a kind of navel orange pol based on EO-1 hyperion through-transmission technique spectrum peak area and detect rapid modeling method, forecast model is set up to navel orange pol, effective extraction hyperspectral information, improves modeling efficiency and accuracy of detection.
The object of the present invention is achieved like this:
Because navel orange pol weighs an important indicator of navel orange quality, transmission method both can make the interior lights spectrum information relevant to navel orange pol effectively be obtained, and don't can cause navel orange internal injury because light source power is too high; Utilize EO-1 hyperion collection of illustrative plates peak area to carry out the differentiation of navel orange pol, the modeling process of less spectral variables complexity, computing velocity is fast, and accuracy rate is higher, can meet the requirement to navel orange pol Fast nondestructive evaluation.
Specifically, this method comprises the following steps:
1. navel orange sample half transmitting EO-1 hyperion collection of illustrative plates is obtained
EO-1 hyperion instrument is adopted to detect the EO-1 hyperion collection of illustrative plates of navel orange sample, setting acquisition mode, time shutter, light source power, wavelength coverage, resolution and picking rate;
2. the pol value of chemical gauging navel orange sample is utilized
The pol value of navel orange sample is measured according to assay method described in GB GB/T8210;
3. navel orange sample average EO-1 hyperion collection of illustrative plates is chosen
Choose navel orange mean height spectrum atlas, according to navel orange high-spectrum spectral property, under MATLAB environment, Pretreated spectra is carried out to the curve of spectrum, after removing non-targeted information, instrument noise, background interference and removing the irrelevant variable information such as water peak, simplify spectral information, retain important information;
4. EO-1 hyperion collection of illustrative plates spectrum peak area is calculated
Under MATLAB environment, matching is carried out to the curve of spectrum, crest value and trough value under self-adaptation chosen spectrum region, by integral and calculating EO-1 hyperion collection of illustrative plates spectrum peak area;
5. set up navel orange sample glucose prediction model, carry out Quality Detection
Utilize the input variable of ratio as model of left and right spectrogram spectrum peak area under navel orange sample self-adaptation chosen spectrum region, pol Quality Detection is carried out to unknown navel orange sample, the quantitative detection model of linear regression is set up to navel orange pol, utilizes forecast set related coefficient and forecast set root-mean-square error to evaluate the precision of detection model.
The present invention has following advantages and good effect:
1. by gathering the half transmitting EO-1 hyperion of navel orange sample, thus the difficult point that the thicker spectroscopic methodology of Peel of Navel Orange is difficult to obtain inside quality can effectively be solved;
2. the correlationship utilizing EO-1 hyperion collection of illustrative plates peak area and navel orange sample pol to change, decreases high-spectral data amount greatly on the impact of model calculation speed;
3. this rapid modeling method can meet the requirement of high speed pachydermia class fruit on-line checkingi, completes and obtains effective spectral signal at short notice and set up regression model, improve detection efficiency and accuracy of detection.
In a word, the present invention is based on hyperspectral technique and gather navel orange sample EO-1 hyperion collection of illustrative plates by half transmitting mode and effectively can obtain navel orange inside quality information, improve detection level and the detection efficiency of fruit internal quality; This modeling modeling efficiency is high, accuracy rate is high, and model calculation speed is fast, can detect the pol Internal quality index of fruit fast, evaluate the inside quality of fruit.
Accompanying drawing explanation
Fig. 1 is the structural representation of this device, in figure:
0-navel orange sample,
1-computer, 2-casing, 3-hyperspectral imager, 4-electricity driving displacement platform,
5-hand-operated lifting platform, 6-light source, 7-sample chamber;
Fig. 2 is the block diagram of this law;
Fig. 3 is the half transmitting averaged spectrum curve of single navel orange sample;
Fig. 4 is the navel orange curve of spectrum after curve;
Fig. 5 is the related coefficient figure of navel orange sample modeling collection glucose prediction value and measured value;
Fig. 6 is the related coefficient figure of navel orange sample forecast set glucose prediction value and measured value.
Embodiment
Describe in detail below in conjunction with drawings and Examples:
One, device
As Fig. 1, this device comprises target---navel orange sample 0;
Be provided with computer 1, casing 2, hyperspectral imager 3, electricity driving displacement platform 4, hand-operated lifting platform 5, light source 6 and sample chamber 7;
Its position and annexation are:
Bottom in casing 2 is provided with electricity driving displacement platform 4, and the top of electricity driving displacement platform 4 is provided with hand-operated lifting platform 5, is provided with light source 6 in sample chamber 7, is placed with navel orange sample 0 on hand-operated lifting platform 5;
Top in casing 2 is provided with hyperspectral imager 3;
Electricity driving displacement platform 4 and hyperspectral imager 3 are connected with the computer 1 outside casing 2 respectively.
Above-mentioned each functional part is general part.
Its working mechanism is: open computer 1 and hyperspectral imager 3, connect the light source 6 of sample chamber 5, preheating 30 minutes, the picking rate and collection position that software are arranged time shutter, wavelength coverage and resolution and electricity driving displacement platform 4 is carried at hyperspectral imager 3, regulate the object lens of hyperspectral imager 3, obtain the image of navel orange sample 0 clearly; Navel orange sample 0 is positioned over hand-operated lifting platform 5, the navel orange sample 0 in adjustment sample chamber 7 and the position of light source 6, click software and start to gather button, electricity driving displacement platform 4 moves, and completes the collection of navel orange sample high spectrum image simultaneously.
Two, method
As Fig. 2, this law comprises the following steps:
1. navel orange sample half transmitting EO-1 hyperion collection of illustrative plates A is obtained
Adopt EO-1 hyperion instrument to detect the EO-1 hyperion collection of illustrative plates of navel orange sample, acquisition mode is half transmitting mode, and spectroscopic light source is 4 50W Halogen lamp LEDs, light source general power is 200W, setting time shutter 100ms, and wavelength coverage is 300 ~ 1100nm, picking rate is 0.2cm/s, and resolution is 32cm
-1; Gather the transmission high-spectrum spectrum information at position, navel orange sample equator, in the middle part of automatic acquisition image, pixel size is the spectral value in the image range of 120 × 120pixels respectively;
2. the pol value B of chemical gauging navel orange sample is utilized
The pol value of navel orange sample is measured according to assay method described in GB GB/T8210;
3. navel orange sample average EO-1 hyperion collection of illustrative plates C is chosen
Choose navel orange sample average EO-1 hyperion collection of illustrative plates, according to navel orange high-spectrum spectral property, in MATLAB software, the level and smooth pre-service of S-G is carried out to spectroscopic data, after removing non-targeted information, instrument noise, background interference and removing the irrelevant variable information such as water peak, simplify spectral information, retain important information;
4. EO-1 hyperion collection of illustrative plates spectrum peak area D is calculated
In MATLAB software, carry out matching to the curve of spectrum, self-adaptation chooses crest value and trough value under 390 ~ 1055nm SPECTRAL REGION, by integral and calculating EO-1 hyperion collection of illustrative plates spectrum peak area;
5. set up navel orange sample glucose prediction model, carry out Quality Detection E
Navel orange sample self-adaptation is utilized to choose the input variable of ratio as model of spectrogram spectrum peak area under 390 ~ 1055nm SPECTRAL REGION, pol Quality Detection is carried out to unknown navel orange sample, the quantitative detection model of linear regression is set up to pol, utilizes forecast set related coefficient and forecast set root-mean-square error to evaluate the precision of detection model.
Three, test findings
For the high spectrum image collection of illustrative plates of navel orange sample, after carrying out integral and calculating by EO-1 hyperion collection of illustrative plates peak area, set up navel orange pol detection model, and model inspection precision is analyzed.
Fig. 2 is the block diagram of this law, and Fig. 3 is the half transmitting averaged spectrum curve of single navel orange sample, and spectral range is 390 ~ 1055nm; Gather navel orange sample high-spectrum spectrum information, acquisition mode is half transmitting mode, and spectroscopic light source is 4 50W Halogen lamp LEDs, and light source general power is 200W, and setting time shutter 100ms, wavelength coverage is 300 ~ 1100nm, and picking rate is 0.2cm/s, and resolution is 32cm
-1; Gather the transmission high-spectrum spectrum information at position, navel orange sample equator, in the middle part of automatic acquisition image, pixel size is the average light spectrum in the image range of 120 × 120pixels respectively; Gathered spectrum, measure navel orange sample pol value according to assay method described in GB GB/T8210, pol value scope is 9.6 ~ 12.6
obrix; Be modeling collection and forecast set by all samples according to the ratio cut partition of 2:1, the pol value of modeling collection sample comprises the pol value of forecast set sample.
Described navel orange pol fast modeling method, be the image of 120 × 120pixels from navel orange high spectrum image centre selected pixels size in ENVI software, and extract the averaged spectrum of every width image, spectral range is 390 ~ 1055nm.
Gained navel orange sample spectra is carried out the level and smooth pre-service of Savitzky-Golay, carries out curve fitting in Matlab software and obtain crest and the wave trough position of spectrum in curve, calculate the peak area of two peak regions in navel orange EO-1 hyperion collection of illustrative plates.
Using calculating the ratio of left and right peak area in gained collection of illustrative plates as the input variable of navel orange glucose prediction, set up linear regression model (LRM); Set up linear model RMSEC and RMSEP and be respectively 0.312
obrix and 0.306
obrix, both relatively, now the related coefficient of modeling collection and forecast set is respectively 0.832 and 0.846.
Fig. 5,6 is the related coefficient figure of linear model watermelon glucose prediction value and measured value; As can be seen from the figure, the linear model built obtains and preferably predicts the outcome.
Claims (3)
1. the navel orange pol based on EO-1 hyperion through-transmission technique spectrum peak area detects a rapid modeling method, it is characterized in that comprising the following steps:
1. navel orange sample half transmitting EO-1 hyperion collection of illustrative plates (A) is obtained
EO-1 hyperion instrument is adopted to detect the EO-1 hyperion collection of illustrative plates of navel orange sample, setting acquisition mode, time shutter, light source power, wavelength coverage, resolution and picking rate;
2. the pol value (B) of chemical gauging navel orange sample is utilized
The pol value of navel orange sample is measured according to assay method described in GB GB/T8210;
3. navel orange sample average EO-1 hyperion collection of illustrative plates (C) is chosen
Choose navel orange mean height spectrum atlas, according to navel orange high-spectrum spectral property, under MATLAB environment, Pretreated spectra is carried out to the curve of spectrum, after removing non-targeted information, instrument noise, background interference and removing the irrelevant variable information such as water peak, simplify spectral information, retain important information;
4. EO-1 hyperion collection of illustrative plates spectrum peak area (D) is calculated
Under MATLAB environment, matching is carried out to the curve of spectrum, crest value and trough value under self-adaptation chosen spectrum region, by integral and calculating EO-1 hyperion collection of illustrative plates spectrum peak area;
5. set up navel orange sample glucose prediction model, carry out Quality Detection (E)
Utilize the input variable of ratio as model of left and right spectrogram spectrum peak area under navel orange sample self-adaptation chosen spectrum region, pol Quality Detection is carried out to unknown navel orange sample, the quantitative detection model of linear regression is set up to navel orange pol, utilizes forecast set related coefficient and forecast set root-mean-square error to evaluate the precision of detection model.
2. detect rapid modeling method by navel orange pol according to claim 1, it is characterized in that described step 1. its device adopted be:
Bottom in casing (2) is provided with electricity driving displacement platform (4), and the top of electricity driving displacement platform (4) is provided with sample chamber (7), is provided with light source (6) in sample chamber (7), is placed with navel orange sample (0) on hand-operated lifting platform (5);
Top in casing (2) is provided with hyperspectral imager (3);
Electricity driving displacement platform (4) is connected with casing (2) computer outward (1) respectively with hyperspectral imager (3).
3. detect rapid modeling method by navel orange pol according to claim 1, it is characterized in that parameter of described step 1. its setting is:
Adopt EO-1 hyperion instrument to detect the EO-1 hyperion collection of illustrative plates of navel orange sample, acquisition mode is half transmitting mode, and spectroscopic light source is 4 50W Halogen lamp LEDs, light source general power is 200W, setting time shutter 100ms, and wavelength coverage is 300 ~ 1100nm, picking rate is 0.2cm/s, and resolution is 32cm
-1; Gather the transmission high-spectrum spectrum information at position, navel orange sample equator, in the middle part of automatic acquisition image, pixel size is the spectral value in the image range of 120 × 120pixels respectively.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510644358.2A CN105158178B (en) | 2015-10-08 | 2015-10-08 | Navel orange pol detection rapid modeling method based on EO-1 hyperion through-transmission technique spectral peak area |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510644358.2A CN105158178B (en) | 2015-10-08 | 2015-10-08 | Navel orange pol detection rapid modeling method based on EO-1 hyperion through-transmission technique spectral peak area |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105158178A true CN105158178A (en) | 2015-12-16 |
| CN105158178B CN105158178B (en) | 2018-04-20 |
Family
ID=54799113
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510644358.2A Expired - Fee Related CN105158178B (en) | 2015-10-08 | 2015-10-08 | Navel orange pol detection rapid modeling method based on EO-1 hyperion through-transmission technique spectral peak area |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105158178B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108846203A (en) * | 2018-06-13 | 2018-11-20 | 深圳市华慧品牌管理有限公司 | The method and device of fruit non-destructive testing |
| CN108872085A (en) * | 2018-05-09 | 2018-11-23 | 南京林业大学 | The rotten heart lossless detection method of navel orange based on the detection of blue wave band transmission image |
| CN111795932A (en) * | 2020-06-15 | 2020-10-20 | 杭州电子科技大学 | Hyperspectrum-based nondestructive testing method for sugar acidity of waxberry fruits |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TW200504347A (en) * | 2003-07-25 | 2005-02-01 | Taoyuan Distr Agricultural Res And Extension Station | Photoelectric fruit sugar degree detecting technique |
| WO2005108956A1 (en) * | 2004-05-10 | 2005-11-17 | Hiroshi Maeda | On-line internal quality examining method and device |
| CN201233362Y (en) * | 2008-06-04 | 2009-05-06 | 江西农业大学 | Multi optical spectrum imaging device for detecting fruit quality |
| CN103472031A (en) * | 2013-09-20 | 2013-12-25 | 华东交通大学 | Navel orange sugar degree detection method based on hyper-spectral imaging technology |
| CN103822879A (en) * | 2014-02-24 | 2014-05-28 | 西北农林科技大学 | Nondestructive detection method of swelled kiwi fruits based on hyperspectral imaging technology |
-
2015
- 2015-10-08 CN CN201510644358.2A patent/CN105158178B/en not_active Expired - Fee Related
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TW200504347A (en) * | 2003-07-25 | 2005-02-01 | Taoyuan Distr Agricultural Res And Extension Station | Photoelectric fruit sugar degree detecting technique |
| WO2005108956A1 (en) * | 2004-05-10 | 2005-11-17 | Hiroshi Maeda | On-line internal quality examining method and device |
| CN201233362Y (en) * | 2008-06-04 | 2009-05-06 | 江西农业大学 | Multi optical spectrum imaging device for detecting fruit quality |
| CN103472031A (en) * | 2013-09-20 | 2013-12-25 | 华东交通大学 | Navel orange sugar degree detection method based on hyper-spectral imaging technology |
| CN103822879A (en) * | 2014-02-24 | 2014-05-28 | 西北农林科技大学 | Nondestructive detection method of swelled kiwi fruits based on hyperspectral imaging technology |
Non-Patent Citations (2)
| Title |
|---|
| 李文典等: "应用光谱峰面积法测定食品微量铁含量", 《食品研究与开发》 * |
| 赵珂等: "基于近红外光谱技术的脐橙快速无损检测", 《激光与红外》 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108872085A (en) * | 2018-05-09 | 2018-11-23 | 南京林业大学 | The rotten heart lossless detection method of navel orange based on the detection of blue wave band transmission image |
| CN108872085B (en) * | 2018-05-09 | 2023-04-25 | 南京林业大学 | Navel orange heart rot nondestructive testing method based on blue wave band transmission image detection |
| CN108846203A (en) * | 2018-06-13 | 2018-11-20 | 深圳市华慧品牌管理有限公司 | The method and device of fruit non-destructive testing |
| CN111795932A (en) * | 2020-06-15 | 2020-10-20 | 杭州电子科技大学 | Hyperspectrum-based nondestructive testing method for sugar acidity of waxberry fruits |
| CN111795932B (en) * | 2020-06-15 | 2022-11-15 | 杭州电子科技大学 | Hyperspectrum-based nondestructive testing method for sugar acidity of waxberry fruits |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105158178B (en) | 2018-04-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103630499B (en) | A kind of fish protein content distribution detection method based on high light spectrum image-forming technology | |
| CN103674864B (en) | A kind of fish water content distribution detection method based on high light spectrum image-forming technology | |
| CN104062263B (en) | The near-infrared universal model detection method of light physical property close fruit quality index | |
| CN106841103A (en) | Near infrared spectrum detects fruit internal quality method and dedicated test system | |
| CN101308086B (en) | Fruit internal quality on-line checking apparatus based on near infrared spectra technology | |
| CN105044021B (en) | A kind of mid-autumn crisp jujube pol lossless detection method | |
| CN106000914A (en) | Apple sorting device for on-line nondestructive test on moldy core | |
| CN108846203A (en) | The method and device of fruit non-destructive testing | |
| CN101413885A (en) | Near-infrared spectrum method for rapidly quantifying honey quality | |
| CN101929951A (en) | Method for distinguishing milk doped with ewe's milk by near infrared spectrum | |
| CN105067531A (en) | Mango quality nondestructive detection method and mango quality nondestructive detection apparatus | |
| CN105092579A (en) | Mango quality non-destructive testing device | |
| CN106841083A (en) | Sesame oil quality detecting method based on near-infrared spectrum technique | |
| CN205762405U (en) | Sorting unit for online Non-Destructive Testing Apple Mould Core equipment | |
| CN102937575B (en) | Watermelon sugar degree rapid modeling method based on secondary spectrum recombination | |
| CN103487397B (en) | A kind of thunder bamboo shoots hardness method for quick and device | |
| CN104280349A (en) | Method for identifying hollowness of white radishes based on hyperspectral image | |
| CN103487422A (en) | Cloud traditional Chinese medicine quality detection system of multi-wavelength LED fluorescence spectrum and method | |
| CN103134770B (en) | Eliminate moisture detects total nitrogen content of soil impact method near infrared spectrum | |
| CN105158178A (en) | Rapid modeling method for detecting sugar content of navel orange based on spectral peak area in high spectral transmission technology | |
| Wang et al. | Development of near‐infrared online grading device for long jujube | |
| CN110376154A (en) | Fruit online test method and system based on spectrum correction | |
| CN206788033U (en) | A kind of dedicated test system of near infrared spectrum detection fruit internal quality | |
| CN105866043A (en) | Method for detecting apple sour through hyperspectral technology | |
| CN203572772U (en) | Cloud traditional Chinese medicine quality detection system based on multi-wavelength LED fluorescence spectrum |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20180420 Termination date: 20181008 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |