JPH0475467B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION (a) Field of Industrial Application The present invention relates to a circuit for detecting a peak position in a spectrum obtained in spectroscopic measurement or a chromatogram in chromatography.

(b) Prior Art Usually, the operating principle of an apparatus for detecting a peak position in a spectrum or chromatogram is to detect a zero-crossing point of a first-order differential of a spectrum or the like.
However, with this method, if a small peak is on top of the slope of a large peak, as shown by 1 in Figure 1, or if the slope of the baseline is large, the first derivative will be calculated as shown by 2 in Figure 1. There is a zero-crossing point for relatively large peaks such as peaks Kuro and C, but zero-crossing points do not appear for small peaks such as A and D, and with this method, the curve is on a slope. Small peaks may not be detected.

(c) Purpose The present invention aims to provide a peak detection circuit that can detect even extremely small peaks on a slope as shown by A and D in FIG.

(d) Structure Differential operation is to divide the differential dy of the dependent variable by the differential dx of the independent variable, but the general method when calculating the differential of measurement data using an instrument is to divide the independent variable by a distance of Δx as dx. 2 points are taken, and the difference Δy of the measured output between them is taken as the differential of y, and Δy/
This is to calculate Δx. In the present invention, two types of intervals, large and small, are prepared for the interval between two points on an independent variable for performing a differential operation, and the interval Δx ₁ of the smaller one is set to a value sufficiently smaller than the width of the peak to be detected, and the interval of the other For example, set the larger Δx2 to the same level or greater than the width of the peak to be detected, and detect the zero-crossing point at the difference between the first-order differential functions of the measured data calculated using the differentials of these independent variables. It is something to do.

Now consider finding the first derivative of the function shown in Figure 2 using Δx1 and Δx2. When Δx1 is used, the maximum value of the first-order differential Δy1/Δx1 is very close to the slope of the tangent a at the maximum slope point of this function.
On the other hand, the maximum value of the first-order differential obtained using Δx2 is the slope of the tangent line b shown in the figure, which is smaller than the slope of the tangent line a. On the other hand, for a slope line as shown in FIG. 3, the first-order differential using either Δx1 or Δx2 has the same value. Therefore, if we consider the case where a low peak A is on a sloped baseline B as shown in Fig. 4, the differential with respect to the baseline is Δx1,
Since the values obtained by both Δx2 are equal, they are eliminated by subtraction. On the other hand, the differential value for the peak component is largely due to Δx1, and the value calculated using Δx2 is 0.
The difference between the two differentials is very close to the first-order differential obtained using Δx1 only for the peak component A excluding the baseline, and it becomes possible to detect the zero-crossing point. The peaks A and D in curve 1 in FIG. 1 correspond to the case in FIG. 4.

(E) Example Since the measured output often contains noise, it is necessary to average the noise when performing differential salt calculation. There are various methods of noise averaging, one of which is the weighted average method. In Figure 5, x1, x2...
xn... is the sampling point of the measured value, and the vertical axis indicates the weight of the measured value. Let us now calculate the average value of the measured values at xn points. 2k+ around point xn
The value weighted averaging method multiplies the data of one sampling point by the value weight of the ratio shown in the vertical bar in the figure to find the average, and uses it as the measured value of xn. This method
We will proceed for each point xn, xn+1, xn+d,... An extension of this idea to differential operations has been proposed by Savitzky and Golay. Similarly to FIG. 5, FIG. 6 also shows the sampling points on the horizontal axis, and the weight applied to the measured value on the vertical axis. When finding the first derivative at xn point,
Take the same number of sampling points before and after the xn point, set the positive and negative points as shown in the figure, and add the multiplied values to the measured value, (sampling interval) x (number of sampling points - 1) Divide by. In this case, what corresponds to Δx mentioned in the previous section is the width indicated by Δx in the figure.

The embodiment described here utilizes the Zavitsky and Golay method. FIG. 7 shows the configuration of the embodiment. MA is the measurement device, SH is the sample and hold circuit for sampling the measurement device output, and AD is the A/
A D converter converts the measured value held by the sample and hold circuit SH into digital data. M is a memory in which the digital data is stored. The CPU controls the entire device,
The arithmetic control circuit that performs the differential and subtraction operations described above, and D is a display device such as a recorder or printer.

The arithmetic control circuit CPU drives the measuring device, operates the sample hold circuit SH and the A/D converter AD, samples the measurement output at regular intervals, and stores it in the memory M. At the same time, the measurement data stored in the memory M is read out, the above-mentioned differential calculation is performed using two types of large and small Δx, the calculation is executed to calculate the difference between the two differential values corresponding to the same measurement point, and the result is Store it in memory M. In order to obtain the differential of a certain point, data of several sampling points before and after the certain point are required, so the above-mentioned calculation proceeds a little later than the progress of the measurement operation. According to the progress of the above calculation, the CPU searches for a zero-crossing point from the subtraction result, outputs the data at the wavelength value and time corresponding to that point to the display device, and displays the measurement output and data regarding each peak on the display device. do.

In the present invention, the differential calculation method is arbitrary. Further, in the above embodiment, a real-time method is adopted in which measurement and calculation are performed substantially simultaneously, but it is also possible to temporarily store all the measurement data in a memory and read it out later to perform the above-mentioned calculation.

(F) Effects According to the present invention, it is possible to sensitively detect a small peak whose sloped baseline is on the slope of a large peak.

[Brief explanation of the drawing]

Figure 1 is a graph explaining the purpose of the present invention, Figure 2 is a graph explaining the purpose of the invention.
The figure is a graph explaining the difference in differential sensitivity, Figure 3 is a graph explaining that the same differential value is obtained regardless of the differential sensitivity for a constant slope, Figure 4 is a graph showing the peak on the slope, FIG. 5 is a graph showing how to apply weights in the weighted average method, FIG. 6 is a graph showing weights for differential calculations, and FIG. 7 is a block diagram of an apparatus according to an embodiment of the present invention. MA...Measuring device, SH...Sample and hold circuit, M...Memory, D...Display device.

Claims (1)

[Claims] 1 Differential calculation means for sampling the measurement circuit output and performing two types of differential calculations, difference calculation means for calculating the difference between the two types of differential calculation values obtained, and control for detecting the zero-crossing point of the difference as the peak position The above two types of differential calculations are based on a circuit in which the value of the differential of the independent variable used in the differential calculation is sufficiently smaller than the width of the peak to be detected for one, and the same as the width of the peak to be detected for the other. 1. A peak detection circuit for a measured output, characterized in that the peak detection circuit has a value equal to or greater than that.