Chinese herbal medicine powder phase spot data acquisition technology

Abstract

The application discloses a data collection technology for Chinese herbal medicine powder phase spot, which realizes the data feature extraction technology device of the powder stacking image of quantitative powder in a medicine bin under the action of an electric field. The herbal medicine powder is quantitatively filled into a feeding opening, and the powder particles fall freely from high to low under the action of a directional electric field after the feeding opening, and then the powder particles fall to a transparent bottom plate, and on the basis of the formed scattering distribution image, the edge feature parameters of the powder spot image are extracted, and the feature data λ is constructed. The feature parameters are formed by a parameter formula generated by the spot image. The collected feature data is composed of 10 λ values arranged according to the size.

Description

Technical Field

[0001] In the field of electronic sensors for traditional Chinese medicine, a device for extracting characteristic data of traditional Chinese medicine powder has been developed. Background Technology

[0002] The general characteristics of the powder state of Chinese herbal medicines have been found. The main methods used are microscopic techniques [1], mechanical sieves (mesh) [2], electric field (dust) [3] and liquid medium oil analysis [4]. However, it is difficult to find the individual characteristics or category characteristics of the corresponding Chinese herbal medicines. The technology of identifying and classifying the powder state of Chinese herbal medicines using image technology [5] and artificial intelligence technology [6] has also been applied to individual varieties. However, it does not have general characteristics. The mechanical funnel powder processing technology [7] and the characteristics of Chinese herbal medicine powder images [8] have made certain suggestions on powder collection and possible powder image characteristics. However, neither of them has proposed a general collection method and technical points. Powder particle size measurement technology is abundant [9]. Generally, particles with a particle size greater than 50 μm settle quickly under the action of gravity. Fine powder is the main component after Chinese herbal medicine is crushed and is distributed in the range of fine particles and fine powder (100-1000 μm)

[10] . It has a light-blocking effect. The equivalent diameter dH of the projected circle, also known as the Heywood diameter, is introduced, which is the diameter of the circle with the same projected area as the particle

[11] . If the projected area of ​​the particle is A, then . [Literature and References] [1] CN101126699B. A pollen particle image data acquisition system. [2] CN211303207U. A traditional Chinese medicine pulverizing and screening device. [3] CN206642854U. An electric field dispersion device for root and tuber powder. [4]CN215876316U. A system for extracting traditional Chinese medicine powder. [5] CN112102318A. A method and apparatus for identifying traditional Chinese medicine based on multimodal image recognition. [6] CN111709389A. Intelligent identification method and system for traditional Chinese medicine powder based on microscopic images. [7] CN211217404U. A novel particulate powder separation device. [8] Amuguleng, Hassurong, Gao Luyan, et al. Digital representation of microscopic features of Chinese herbal medicine powders [C] / / Proceedings and abstracts of the 9th Academic Conference of the Veterinary Pharmacology and Toxicology Branch of the Chinese Association of Animal Science and Veterinary Medicine. 2006. DOI:ConferenceArticle / 5aa2d384c095d72220a9e446. [9] Wu Li, Wang Xiaowei, Lu Xingjie, Zhu Yonghong. Current status and application progress of particle testing technology [J]. Industrial Metrology, 2019, 29(01):1-8. DOI:10.13228 / j.boyuan.issn1002-1183.2018.0148.

[10] Chinese Pharmacopoeia (2020 edition).

[11] Hao Jiming, Ma Guangda, Wang Shuxiao. Air Pollution Control Engineering (3rd Edition) [M]. Higher Education Press. January 15, 2010. Summary of the Invention

[0003] This invention realizes a feature data extraction technology for obtaining a distribution image of scattered medicinal herb powder particles on a transparent test plate under the action of an electric field. The medicinal powder enters the test chamber A10 through the feeding port A20 and is guided to the test plate A30 by an electric field generated by an electric field generator A40, quantitatively dispersing the powder onto the transparent test plate. Then, it is illuminated by LED light sources A11 and A12 fixed inside the chamber, causing the powder to disperse on the test plate and present an image of its distribution. A camera circuit A50 on the other side of the test plate A30 completes the acquisition of this scattered powder image and data generation. Figure 1 , The resulting image of the medicinal powder is formed by LEDs on a transparent, graduated test plate A30, creating a distribution image. The pulverized medicinal material is loaded into hopper A20, and the upper cover plate A21 is closed. The lower baffle plate A22 is opened, and the powder falls freely onto the dust test plate A30. At the same time, the high-voltage static voltage generator A40 is turned on to supply power to the electrostatic nets A41 and A42 at the bottom for 10 seconds and then turned off. Subsequently, LED lights A11 and A12 are turned on, and the image acquisition circuit (camera) A50 is activated to acquire dust (powder) images. The corresponding circuit device consists of a switch on the upper cover plate A21 and a switch on the lower baffle plate A22, initiating the powder feeding preparation process. After a 5-second timeout, the lower baffle plate A22 opens, allowing the medicinal powder to fall. The timer is then restarted for another 5 seconds. The powder freely falls onto the test plate A30 and remains there for 5 seconds. At the end of this 5-second timeout, the electrostatic generator A40 is activated, supplying power to the electrostatic grid electrodes A41 and A42, accelerating the directional fall of the floating dust onto the test plate A30. This electrostatic process lasts for 10 seconds. Afterward, the LED lights are activated, and the camera captures a photo of the transparent substrate. Its logic framework is as follows: Figure 2 , The distribution characteristics of dust can be determined by the light transparency of the transmitted powder spot. For example... Figure 3 , The distribution of dust (powder) from the Chinese medicine powder falling onto a transparent plate, forming a central opaque spot that gradually becomes transparent. Figure 4The grayscale value γ at the edge of the medicinal dust spot is related to the individual shape, weight, and dielectric properties of different dust particles. By calculating the grayscale value change Δγ at the edge of the dust spot and the corresponding distance τ, the ratio λ between the two is given and defined as the characteristic value λ of this dust spot. The calculation formula is as follows: (1) The corresponding pseudocode calculation logic is as follows: The logic for generating the feature dataset of the target Chinese herbal medicine based on the powder distribution image obtained on the A30 test plate is as follows: Acquire powder images Grayscale processing Determine the threshold S0, and obtain the spot features in the image based on the pixel adjacency algorithm. Output the current segment λ ; Traverse all spot values ​​λ Sort and select the 10 largest values. ;Finish The resulting 10 λ sorting values ​​are called the feature value data of the Chinese medicine powder in the object. Detailed Implementation

[0004] 1. The process of dust settling and stacking of medicinal herbs The target medicinal powder has a particle size between 10 micrometers and 0.1 millimeters. When light shines on the powder, a complex transmission process occurs, including diffraction (interference) at the 10-micrometer level, multiple reflections (at the 0.1-millimeter level), and direct light (unobstructed light). Simultaneously, the target medicinal particles also possess various dielectric properties, exhibiting sensitivity to electrostatic environments. When the particles fall onto the test plate under the influence of an electric field, a powder stacking image is formed. Figure 5 ), The powder (spots) formed by dust accumulation, using the Heywood diameter and the actual projected occlusion area, under light illumination, results in transmission and complete occlusion [1, 2] conclusions. [1] Hou, Xiaohong. Research on Coal Powder Particle Detection Technology Based on Image Processing [D]. Master's Thesis. Taiyuan University of Technology. 2006.5 [2] Jin, Meihua. Numerical Study on the Influence of Particle Geometric Properties on the Dispersion Characteristics of Aluminum Powder in 20L Spheres [D]. Tianjin University of Technology. 2023. In summary, the distance of dust projection can be used as a good substitute for area calculation, thus simplifying the calculation of occlusion area and making the application of the transmittance model more specific. 2. Formula for calculating characteristic parameters This is a formula for extracting powder feature data based on a powder image. For example... Figure 5The image shows a stacked powder image process model. Based on the light-blocking characteristics of the particles, there is a clear linear region between sparse and dense particle distributions and the intensity of transmitted light. Calculating the transmittance based on the projected circle is a process related to the particle diameter *r*. When the sum of all particle diameters equals the particle distribution range (scale), an opaque region is considered to exist. When the sum of all particle diameters is less than the particle distribution range (scale), a translucent region is considered to exist. Assume there exists a light intensity γ, corresponding to the perceptible light intensity Δγ in the opaque region; and simultaneously measure the distance τ from the corresponding transparent region to the opaque region. Then define the formula: It is a measurable characteristic parameter of the powder particle size of the reaction object in the light-transmitting area. The light intensity γ of a powder image refers to the grayscale value of the grayscale image after algorithm processing, which is directly obtained from image grayscale conversion. The transmittance of a powder image refers to the grayscale value 's' portion of the grayscale image after algorithm processing. The definition of opacity is determined by the grayscale value 's'. λ , greater than s λ This is the opaque position. Where τ is defined as the value from the minimum grayscale value (0) to s in a continuous grayscale range. λ distance. Corresponding calculation formula: 3. Image grayscale processing formula The commonly used maximum value grayscale calculation method involves taking the maximum value from the three color channels (Red, Green, Blue) of each pixel in the image and using it as the grayscale value of that pixel. Gray = max(Red, Green, Blue) For each image of the target medicinal powder, the pixels are traversed, and the image is processed using the following formula to output a grayscale image. 4. Feature value τ and brightness γ of dust image on drop test plate (collected data) The logic for obtaining the edge feature parameter τ of the powder spot and the brightness γ (i.e., gray level S). For example... Figure 6 As shown, Where τ represents the maximum edge value of the same spot corresponding to the two grayscale thresholds Sλ and S0. The determination of the same spot uses a pixel adjacency algorithm; the brightness γ is secondly represented by the grayscale value S, and Sλ = τ1, S0 = τ0. 5. Characteristic value λ group of multiple powder spots The image contains n dust spots, each corresponding to a portion of the light-transmitting area at its edge. Pseudo-process logic: Step 1: Image grayscale conversion Step 2: Algorithm for detecting adjacent spots larger than the threshold Step 3: Generate Spot ID k Assignment Step 4: Assign grayscale value S λ (=160), minimum value S0 (=0) Step 5: Detect the maximum distance τ on each spot based on spot ID. Step 6: Calculate the formula λ=s λ / τ Step 7: Iterate through all IDs k Obtain the array {λ1, λ2, ..., λ} k} Step 8: Extract the top ten groups with the highest values ​​as output feature data, and pad with zeros if there are fewer than ten groups. 6. Image acquisition and processing circuit An image processing and control circuit composed of an embedded processor. It includes components for capturing images, controlling the device's switches, and activating the high-voltage electric field generator circuit. (Example) Figure 7 As shown, The processor circuit uses internal timers to complete the timing tasks and corresponding actions of Timer 1, Timer 2, and Timer 3, and is responsible for activating the electrostatic generator module, processing image data, and generating feature data. 7. High-voltage electric field circuit The electrostatic generator module A40 consists of a switching high-power amplifier circuit and a boost circuit. The PWM pulse signal from the processor is boosted into a high voltage by the boost circuit and applied to plates A42 and A41. Figure 8 As shown, The PWM signal output in the processor is implemented by activating the electrostatic discharge module, and a piece of code that generates the PWM signal is embedded within it. The pseudocode C language logic of the program is as follows: #include<stdio.h> / / Assume the following functions are APIs provided by the PWM library. void pwm_init(unsigned int frequency, unsigned int dutyCycle); void pwm_start(); void pwm_stop(); int main() { / / Configure the PWM output frequency to FHz and the duty cycle to 50%. pwm_init(F, 50); / / Start PWM signal pwm_start(); ... return 0; } Attached Figure Description

[0005] Figure 1 Framework diagram for collecting data on the biomarker images of Chinese herbal medicine powders. Figure 2 Logical framework diagram for generating images of herbal powder spots in objects. Figure 3 A diagram illustrating dust spots falling onto a transparent panel. Figure 4 Actual dust distribution map. Figure 5 A schematic diagram illustrating the calculation of characteristic parameters of the edge of dust accumulation spots. Figure 6 Logic diagram of the algorithm for obtaining the characteristic parameters of powder spots. Figure 7 Image acquisition and processing circuit diagram. Figure 8 A schematic diagram of the high-voltage electrostatic field generation model and circuit structure.

Claims

1. The "Data Acquisition Technology of Chinese Herbal Medicine Powder" realizes the extraction technology device that realizes the extraction of data features from the image by directionally falling a quantitative amount of Chinese herbal medicine powder in a medicine chamber under the action of an electric field and forming a powder stack image on a transparent test plate.

2. The data feature extraction device according to claim 1 includes the feeding port A20, medicine chamber A10, LED image lighting lamp A11 (A12) and electrostatic electrodes A41 and A42 as shown in FIG1.

3. The method described in claim 1, which realizes the directional falling of a quantitative amount of Chinese herbal medicine powder in a medicine chamber under the action of an electric field, forming a powder stacking image on a transparent test plate, includes a logical framework for generating the powder spot image of the Chinese herbal medicine under the combined action of the powder falling freely from the feeding port A20 to the test plate A30 and the electrostatic field (Figure 2).

4. The method for forming a powder stack image on a transparent test plate according to claim 1 includes illumination from positions A12 and A11 inside the drug chamber and acquisition of the object powder image position from position A50 outside the drug chamber.

5. The method for obtaining data features from an image according to claim 1 includes the feature calculation formula (1).

6. The data feature extraction technique according to claim 1 includes the logic of generating a feature dataset of the target Chinese medicine by calculating 10 λ values ​​in the target medicinal powder image.

Images