JPH06201673A - Method for detecting peak of chromatogram - Google Patents

Abstract

PURPOSE:To detect the starting point, vertex, and ending point of a peak by one time of scanning by fixing a noticeable point on chromatogram data and calculating whether the measured values in a fixed section after the noticeable point are larger or smaller than the data at the noticeable point, and then, recognizing a point of inflection by moving the noticeable point to the maximum value in the section when the measured values are larger or to the minimum value in the section when the measured values are smaller. CONSTITUTION:A section width W and continuous section value (n) are decided as peak detecting references. The section width W is set at 1/4 of the width at the half maximum to be detected. After resetting the minimum section value, maximum section value, and mean section value of a memory area, the measured values in the section corresponding to the section width W from a noticeable point are found and, when the mean value of the section is larger than the data at the noticeable point, the section is discriminated as a rising section and, when smaller, as a falling section. When, for example, the noticeable point moves from P0 to P2 through P1 and the means value of the section W starting from P2 is smaller than the data at P2, and then, the position of the minimum value in the section is PS, the noticeable point is moved to PS and, when the peak again starts to rise, the point of inflection PS is discriminated as the starting point of the peak and recorded.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a peak detection method for finding a starting point, an apex and an ending point of a peak of a chromatogram in a data processing device in a liquid chromatograph or a gas chromatograph.

[0002]

2. Description of the Related Art As a method of detecting the position of a peak from chromatogram data, the sampling data is sampled at regular intervals and compared sequentially, and the peak of the peak changes at the point where the data changes from increasing to decreasing. To detect. When the peak apex is found, the data is scanned again around the peak apex to calculate the amount of change in the sampling data from the peak apex one by one, and the point where the specified condition is satisfied is the start or end point of the peak. Is detected as.

[0003]

When a peak is to be detected by the conventional method, it is necessary to scan the chromatogram data twice, which causes a problem that the processing time becomes long.
Further, the start point and the end point of the peak are obtained by the peak detection reference, but the detection reference needs to be changed depending on the sharpness of the peak. When the sharp peaks and the blunt peaks are mixed, it may not be possible to detect any of the peaks unless the peak detection standard is accurately defined.

The present invention makes it possible to detect the start point, apex and end point of a peak by scanning the chromatogram data only once, thereby shortening the processing time and in a wide range from a sharp peak to a dull peak. It is an object of the present invention to provide a method that enables peak detection over a period of time.

[0005]

In the present invention, a point of interest is set on the chromatogram data, and whether the data within a certain section after that point is rising or falling with respect to the point of interest is calculated, If it is, the point of interest is moved to the maximum value in the section, and if it is descending, it is moved to the minimum value in the section to catch the inflection point of the peak and extract the peak. Therefore, in the present invention, a fixed section width W including a plurality of sampling data for chromatogram data and a continuous section value n for judging rise or fall of the data are set, and a point of interest is set on the chromatogram. From the data within a certain section width, the section minimum value, the section maximum value, and the section average value are obtained.If the section average value is larger than the value of the point of interest, the section is raised, and if it is smaller, it is determined to be descending. The point of inflection is detected by moving the point of interest to the point of maximum value in the section and moving the point of interest to the point of minimum value in the section in the descending section, and after descending continuously for n sections or more. The inflection point is defined as the peak start point or the peak end point, and the inflection point after continuously rising for n sections or more is defined as the peak apex.

[0006]

EXAMPLES Examples will be described with reference to FIGS. 1 to 3. Figure 1
2 and FIG. 2 show the procedure of one embodiment in PAD
iagram). As shown in FIG. 1, a section width W and a continuous section value n for determining whether the section is an ascending section or a descending section are determined as peak detection criteria. These can be obtained, for example, as follows based on the noise cycle and the minimum peak to be detected. W = (value less than or equal to ¼ of half-width of peak to be detected) n = (noise period) / (sampling rate) value By setting both W and n small, noise is also detected as a minute peak Then, after that, by adopting a method of including peaks having a small peak area in continuous peaks on the left and right, it is possible to detect peaks with high accuracy.
As an example, W is set to a section corresponding to the number of data 20 and n is set to 4.

Memory areas for (current position), (peak minimum position), (peak maximum position), (section minimum value), (section maximum value) and (section average value) are prepared. As an initial state, the data is assumed to be in a descending state for n consecutive times, the first data is recorded at (current position), and the recording of (peak minimum position) and (peak maximum position) in the memory area is erased.

In the section from (current position) to (current position + section width W) after (section minimum value), (section maximum value) and (section average value) of the memory area are reset,
A section minimum value, a section maximum value, and a section average value are obtained and recorded in each memory area. If the section average is larger than the data of the current position, the section is judged to be an ascending section, and if it is smaller, the section is judged to be a descending section. In case of ascending, it is processed according to FIG. 2A, and in case of descending, it is processed according to FIG. 2B.

The case of detecting the peak in FIG. 3 will be described as an example. Point of interest (the current position) been moved P _0, P _1, P _2, interval average value in a section starting from P ₂ is smaller than the data of the current position P _2, the position of the minimum value of the interval P
s. In the next section, the point of interest (current position) moves to Ps. In the section starting from Ps, assuming that the section average value becomes larger than Ps and starts to rise, FIG.
It is processed according to the procedure of (A). In this case, since neither (peak minimum position) nor (peak maximum position) has been recorded yet, after (current position) Ps is recorded in (peak minimum position), (current position) shows (section maximum value). Position P ₃
Is recorded, and the current position moves to the position P _{3 having the} maximum value in the section. As a result, the inflection point Ps is captured.

After the memory area (section minimum value), (section maximum value) and (section average value) are reset again, in the section from (current position) P ₃ to (current position P ₃ + section width W) , Section minimum value, section maximum value and section average value are obtained and recorded in respective memory areas. As long as the climb section continues, (current position) is the (section maximum value) of that section.
The process is repeated while moving to the position of.

At the rising section including the peak Pt, the point of interest (current position) moves to Pt, and the next section becomes a section having a section width W starting from (current position) Pt. In that section, it starts to descend and is processed according to FIG.
(Current position) Pt is recorded in (peak maximum position).
As a result, the inflection point Pt is captured. After that, the position of (section minimum value) of the section is recorded in (current position), and the current position moves to the position of the section minimum value. Then, as long as the descending section continues, the process is repeated while the (current position) moves to the (section minimum value) position of the section.

When a descending section including Pe is finally reached, the point of interest (current position) moves to Pe, and the next section becomes a section having a section width W starting from (current position) Pe. In that section, it starts to rise and is processed according to FIG. In this case, since the condition that (peak maximum position) is recorded and (peak minimum position) is before (peak maximum position) is satisfied, (peak maximum position) is the peak apex and (peak minimum position). ) Is regarded as the peak starting point and (current position) is the peak ending point, and the peak data is created.

After that, the record of (peak maximum position) is erased, (current position) Pe is recorded in (peak minimum position), and (section maximum value) of that section is recorded in (current position). Then, the current position moves to the position having the maximum value in the section, and the data processing of the next section is started. The end point Pe of the previous peak becomes the start point Ps of the next peak, and the peak detection is repeated in the same manner. Under this condition, the half width is about 46
It was possible to detect two peaks having different times with the same interval width.

[0014]

According to the present invention, since the start point, the peak and the end point of the peak can be detected by scanning the chromatogram data once, the data processing time becomes short.
It is possible to detect peaks without accurately defining the peak detection reference, and it is possible to deal with a wide range of peaks having different half widths.

[Brief description of drawings]

FIG. 1 is a PAD showing an overall peak detection process according to an embodiment.
It is a figure.

FIG. 2 is a PAD diagram showing processing in an ascending section (A) and a descending section (B).

FIG. 3 is a chromatogram peak diagram showing the operation of one example.

Claims (1)

[Claims] 1. A constant section width W containing a plurality of sampling data for chromatogram data, a continuous section value n for judging rise or fall of the data, and an attention point is set on the chromatogram. From the data within the fixed section width W, the section minimum value, the section maximum value and the section average value are obtained. If the section average value is larger than the value of the point of interest, it is determined that the section is up Then, move the point of interest to the point of maximum value in that section, and in the descending section, move the point of interest to the point of minimum value in that section to capture the inflection point of the peak, and after descending continuously for n sections or more The peak detection method is characterized in that the inflection point is defined as the peak start point or the peak end point, and the inflection point after continuously rising for n sections or more is defined as the peak apex.