Intelligent steel classification system based on confidence comparison

Abstract

The present invention relates to an intelligent steel classification system using confidence comparison, comprising: a classification transmission device, used to transmit the target steel to be classified to the top of the visual acquisition station, wherein the target steel to be classified is one of round steel, square steel, I-beam steel, unequal angle steel and hexagonal steel; a directional analysis mechanism, used to use a trained feedforward neural network to intelligently identify the confidence of a pattern to be compared corresponding to the steel classification in the layer-by-layer conversion image based on various visual information. The intelligent steel classification system using confidence comparison of the present invention is intelligent in design and easy to operate. Since the confidence of the pattern to be compared corresponding to the steel classification in the layer-by-layer conversion image can be intelligently identified by using various visual information of the layer-by-layer conversion image, the pattern to be compared is one of round, square, I-beam, unequal angle triangle and hexagon, thereby completing the intelligent classification processing of the target steel.

Description

Technical Field

[0001] The present invention relates to the field of steel classification, and in particular to an intelligent steel classification system using confidence comparison. Background Art

[0002] Steels are classified by forming method as follows: forged steel, cast steel, hot-rolled steel, cold-drawn steel. Steels are classified by metallographic structure as follows: annealed steels include hypoeutectoid steel (ferrite + pearlite), eutectoid steel (pearlite), hypereutectoid steel (pearlite + cementite), ledeburite steel (pearlite + cementite); normalized steels include pearlite steel, bainite steel, martensite steel, austenite steel; steels with no phase change or partial phase change.

[0003] Steel is classified by use as follows: Construction and engineering steel, which includes ordinary carbon structural steel, low alloy structural steel, and reinforced steel; Steel structural steel, which includes machinery manufacturing steel, spring steel, and bearing steel. Among them, machinery manufacturing steel includes quenched and tempered structural steel, surface hardened structural steel (carburizing steel, ammonia impregnated steel, surface hardening steel), easy-cut structural steel, and cold plastic forming steel (cold stamping steel, cold heading steel); Tool steel, which includes carbon tool steel, alloy tool steel, and high-speed tool steel; Special performance steel, which includes stainless acid-resistant steel, heat-resistant steel (oxidation-resistant steel, heat-resistant steel, valve steel), electric heating alloy steel, wear-resistant steel, low-temperature steel, and electrical steel; Special steel, which includes bridge steel, ship steel, boiler steel, pressure vessel steel, agricultural machinery steel, etc.

[0004] Due to the different modes of steel classification, there are many types of classification in each classification mode. Therefore, the manual classification mode is slow and inefficient, with a low level of automation, and the classification accuracy is insufficient, and the classification results are not accurate enough. However, the existing technology lacks a targeted and high-precision steel classification mechanism, which makes it difficult to automatically implement steel classification in the existing technology. Summary of the invention

[0005] In order to solve the technical problems in the related fields, the present invention provides a steel intelligent classification system using confidence comparison, which can parse the grayscale values ​​corresponding to the pixels of the layer-by-layer converted image and the coordinate data corresponding to the pixels of the layer-by-layer converted image, obtain the grayscale values ​​corresponding to the pixels of the pattern to be compared and the coordinate data corresponding to the pixels of the pattern to be compared, and use a trained feedforward neural network to classify the steel intelligent classification system based on the grayscale values ​​corresponding to the pixels of the layer-by-layer converted image, the coordinate data corresponding to the pixels of the layer-by-layer converted image, and the pixel data of the pattern to be compared. The confidence of the pattern to be compared in the layer-by-layer converted image is intelligently identified by using the grayscale values ​​corresponding to each of the pixels of the pattern to be compared, thereby realizing intelligent judgment on whether the pattern to be compared exists in the layer-by-layer converted image, wherein the pattern to be compared is one of a circle, a square, an I-shape, an unequal-sided triangle and a hexagon, and the patterns to be compared are respectively subjected to corresponding intelligent identification processing using the trained feedforward neural network to obtain respective confidences, and the steel type corresponding to the pattern to be compared with the highest confidence is used as the current classification type of the target steel to be classified, thereby completing the intelligent classification processing of the target steel.

[0006] According to the present invention, there is provided a steel intelligent classification system using confidence comparison, the system comprising:

[0007] Classification and transmission equipment, used for transmitting the target steel to be classified to the top of the visual collection station, wherein the target steel to be classified is one of round steel, square steel, I-beam, unequal angle steel and hexagonal steel;

[0008] A network sensing device connected to the classification transmission device is used to perform a photoelectric sensing action on the cross section of the target steel material to be classified based on a network control instruction after detecting that the classification transmission device has completed transmission, so as to obtain and output a corresponding real-time sensing image;

[0009] A layer-by-layer conversion device, connected to the network sensing device and comprising an affine transformation component, a filtering execution component and a data sharpening component, for continuously performing affine transformation processing, salt and pepper noise removal processing and spatial differential method sharpening processing on the received real-time sensing image to obtain and output a corresponding layer-by-layer conversion image;

[0010] A directional parsing mechanism and the layer-by-layer conversion device are used to receive the layer-by-layer conversion image, parse the grayscale values ​​corresponding to the pixels of the layer-by-layer conversion image and the coordinate data corresponding to the pixels of the layer-by-layer conversion image, obtain the grayscale values ​​corresponding to the pixels of the pattern to be compared and the coordinate data corresponding to the pixels of the pattern to be compared, and use a trained feedforward neural network to intelligently identify the confidence of the pattern to be compared in the layer-by-layer conversion image based on the grayscale values ​​corresponding to the pixels of the layer-by-layer conversion image, the coordinate data corresponding to the pixels of the layer-by-layer conversion image, the grayscale values ​​corresponding to the pixels of the pattern to be compared and the coordinate data corresponding to the pixels of the pattern to be compared;

[0011] Wherein, the confidence of the pattern to be compared in the layer-by-layer conversion image is intelligently identified by using a trained feedforward neural network based on each grayscale value corresponding to each pixel point of the layer-by-layer conversion image, each coordinate data corresponding to each pixel point of the layer-by-layer conversion image, each grayscale value corresponding to each pixel point of the pattern to be compared, and each coordinate data corresponding to each pixel point of the pattern to be compared, including: the pattern to be compared is one of a circle, a square, an I-shaped, an unequal-sided triangle and a hexagon, and the patterns to be compared are respectively subjected to corresponding intelligent identification processing by applying the trained feedforward neural network to obtain each confidence, and the steel type corresponding to the pattern to be compared with the highest confidence is used as the current classification type of the target steel to be classified;

[0012] Among them, using a trained feedforward neural network to intelligently identify the confidence of the existence of the pattern to be compared in the layer-by-layer converted image based on each grayscale value corresponding to each pixel point of the layer-by-layer converted image, each coordinate data corresponding to each pixel point of the layer-by-layer converted image, each grayscale value corresponding to each pixel point of the pattern to be compared, and each coordinate data corresponding to each pixel point of the pattern to be compared also includes: the number of times the feedforward neural network is trained is positively correlated with the resolution of the layer-by-layer converted image.

[0013] Therefore, the present invention has at least the following beneficial technical effects:

[0014] Technical effect 1: Analyze the grayscale values ​​corresponding to each pixel of the layer-by-layer conversion image and the coordinate data corresponding to each pixel of the layer-by-layer conversion image, obtain the grayscale values ​​corresponding to each pixel of the pattern to be compared and the coordinate data corresponding to each pixel of the pattern to be compared, so as to provide various basic information for subsequent intelligent identification;

[0015] Technical effect 2: The trained feedforward neural network is used to intelligently identify the confidence of the pattern to be compared in the layer-by-layer converted image based on the grayscale values ​​corresponding to the pixels of the layer-by-layer converted image, the coordinate data corresponding to the pixels of the layer-by-layer converted image, the grayscale values ​​corresponding to the pixels of the pattern to be compared, and the coordinate data corresponding to the pixels of the pattern to be compared, thereby realizing intelligent judgment on whether the pattern to be compared exists in the layer-by-layer converted image;

[0016] Technical effect three: The pattern to be compared is one of a circle, a square, an I-shape, an unequal triangle and a hexagon. The patterns to be compared are respectively subjected to corresponding intelligent identification processing using the trained feedforward neural network to obtain various confidence levels. The steel type corresponding to the pattern to be compared with the highest confidence level is used as the current classification type of the target steel to be classified, thereby completing the intelligent classification processing of the target steel.

[0017] The steel intelligent classification system using confidence comparison of the present invention is intelligently designed and easy to operate. Since the confidence of the pattern to be compared corresponding to the steel classification in the layer-by-layer converted image can be intelligently identified by using various visual information of the layer-by-layer converted image, the pattern to be compared is one of a circle, a square, an I-shaped, an unequal-sided triangle and a hexagon, thereby completing the intelligent classification processing of the target steel. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The embodiments of the present invention will be described below with reference to the accompanying drawings, wherein:

[0019] Figure 1 Schematic diagram of the internal structure of a steel intelligent classification system using confidence comparison according to Embodiment A of the present invention.

[0020] Figure 2 Schematic diagram of the internal structure of a steel intelligent classification system using confidence comparison according to Embodiment B of the present invention.

[0021] Figure 3 Schematic diagram of the internal structure of a steel intelligent classification system using confidence comparison according to Embodiment C of the present invention. DETAILED DESCRIPTION

[0022] The following will describe in detail an embodiment of the intelligent steel classification system using confidence comparison of the present invention with reference to the accompanying drawings.

[0023] Figure 1 This is a schematic diagram of the internal structure of a steel intelligent classification system using confidence comparison according to Embodiment A of the present invention, wherein the system comprises:

[0024] Classification and transmission equipment, used for transmitting the target steel to be classified to the top of the visual collection station, wherein the target steel to be classified is one of round steel, square steel, I-beam, unequal angle steel and hexagonal steel;

[0025] A network sensing device connected to the classification transmission device is used to perform a photoelectric sensing action on the cross section of the target steel material to be classified based on a network control instruction after detecting that the classification transmission device has completed transmission, so as to obtain and output a corresponding real-time sensing image;

[0026] A layer-by-layer conversion device, connected to the network sensing device and comprising an affine transformation component, a filtering execution component and a data sharpening component, for continuously performing affine transformation processing, salt and pepper noise removal processing and spatial differential method sharpening processing on the received real-time sensing image to obtain and output a corresponding layer-by-layer conversion image;

[0027] For example, the layer-by-layer conversion device is connected to the network sensing device and includes an affine transformation component, a filter execution component and a data sharpening component, and is used to continuously perform affine transformation processing, salt and pepper noise removal processing and spatial differential method sharpening processing on the received real-time sensing image to obtain and output the corresponding layer-by-layer conversion image, including: using a PLC device to implement the layer-by-layer conversion device, which is connected to the network sensing device and includes an affine transformation component, a filter execution component and a data sharpening component, and is used to continuously perform affine transformation processing, salt and pepper noise removal processing and spatial differential method sharpening processing on the received real-time sensing image to obtain and output the corresponding layer-by-layer conversion image;

[0028] A directional parsing mechanism and the layer-by-layer conversion device are used to receive the layer-by-layer conversion image, parse the grayscale values ​​corresponding to the pixels of the layer-by-layer conversion image and the coordinate data corresponding to the pixels of the layer-by-layer conversion image, obtain the grayscale values ​​corresponding to the pixels of the pattern to be compared and the coordinate data corresponding to the pixels of the pattern to be compared, and use a trained feedforward neural network to intelligently identify the confidence of the pattern to be compared in the layer-by-layer conversion image based on the grayscale values ​​corresponding to the pixels of the layer-by-layer conversion image, the coordinate data corresponding to the pixels of the layer-by-layer conversion image, the grayscale values ​​corresponding to the pixels of the pattern to be compared and the coordinate data corresponding to the pixels of the pattern to be compared;

[0029] Wherein, the confidence of the pattern to be compared in the layer-by-layer conversion image is intelligently identified by using a trained feedforward neural network based on each grayscale value corresponding to each pixel point of the layer-by-layer conversion image, each coordinate data corresponding to each pixel point of the layer-by-layer conversion image, each grayscale value corresponding to each pixel point of the pattern to be compared, and each coordinate data corresponding to each pixel point of the pattern to be compared, including: the pattern to be compared is one of a circle, a square, an I-shaped, an unequal-sided triangle and a hexagon, and the patterns to be compared are respectively subjected to corresponding intelligent identification processing by applying the trained feedforward neural network to obtain each confidence, and the steel type corresponding to the pattern to be compared with the highest confidence is used as the current classification type of the target steel to be classified;

[0030] Wherein, using a trained feedforward neural network to intelligently identify the confidence of the existence of the pattern to be compared in the layer-by-layer converted image based on each grayscale value corresponding to each pixel point of the layer-by-layer converted image, each coordinate data corresponding to each pixel point of the layer-by-layer converted image, each grayscale value corresponding to each pixel point of the pattern to be compared, and each coordinate data corresponding to each pixel point of the pattern to be compared also includes: the number of times the feedforward neural network is trained is positively correlated with the resolution of the layer-by-layer converted image;

[0031] Among them, using a trained feedforward neural network to intelligently identify the confidence of the existence of the pattern to be compared in the layer-by-layer converted image based on each grayscale value corresponding to each pixel point of the layer-by-layer converted image, each coordinate data corresponding to each pixel point of the layer-by-layer converted image, each grayscale value corresponding to each pixel point of the pattern to be compared, and each coordinate data corresponding to each pixel point of the pattern to be compared also includes: simultaneously inputting each grayscale value corresponding to each pixel point of the layer-by-layer converted image, each coordinate data corresponding to each pixel point of the layer-by-layer converted image, each grayscale value corresponding to each pixel point of the pattern to be compared, and each coordinate data corresponding to each pixel point of the pattern to be compared into the trained feedforward neural network.

[0032] Figure 2 Schematic diagram of the internal structure of a steel intelligent classification system using confidence comparison according to Embodiment B of the present invention.

[0033] and Figure 1 different, Figure 2 The steel intelligent classification system based on confidence comparison can also include the following components:

[0034] a content enhancement mechanism, connected to the directional parsing mechanism, the affine transformation component, the filter execution component and the data sharpening component, respectively, for performing content enhancement on the respective output image signals of the directional parsing mechanism, the affine transformation component, the filter execution component and the data sharpening component, so as to obtain respective content enhanced images corresponding to the respective output image signals of the directional parsing mechanism, the affine transformation component, the filter execution component and the data sharpening component;

[0035] The content enhancement is performed on the output image signals of the directional parsing mechanism, the affine transformation component, the filtering execution component and the data sharpening component to obtain the content enhanced images corresponding to the output image signals of the directional parsing mechanism, the affine transformation component, the filtering execution component and the data sharpening component respectively, including: the first level content enhancement performs geometric correction actions on the output image signals of the directional parsing mechanism, the affine transformation component, the filtering execution component and the data sharpening component respectively;

[0036] Wherein, content enhancement is performed on the output image signals of the directional parsing mechanism, the affine transformation component, the filtering execution component and the data sharpening component respectively to obtain respective content enhanced images corresponding to the output image signals of the directional parsing mechanism, the affine transformation component, the filtering execution component and the data sharpening component respectively, and further comprises: second-level content enhancement performs morphological processing actions on the output image signals of the directional parsing mechanism, the affine transformation component, the filtering execution component and the data sharpening component respectively;

[0037] Wherein, content enhancement is performed on the output image signals of the directional parsing mechanism, the affine transformation component, the filtering execution component and the data sharpening component respectively to obtain respective content enhanced images corresponding to the output image signals of the directional parsing mechanism, the affine transformation component, the filtering execution component and the data sharpening component respectively, and further comprises: the third level content enhancement performs adaptive recursive filtering actions on the output image signals of the directional parsing mechanism, the affine transformation component, the filtering execution component and the data sharpening component respectively;

[0038] Among them, content enhancement is performed on the respective output image signals of the directional parsing mechanism, the affine transformation component, the filtering execution component and the data sharpening component to obtain content enhanced images corresponding to the respective output image signals of the directional parsing mechanism, the affine transformation component, the filtering execution component and the data sharpening component, and also includes: the fourth level content enhancement performs image enhancement actions based on exponential transformation on the respective output image signals of the directional parsing mechanism, the affine transformation component, the filtering execution component and the data sharpening component.

[0039] Figure 3 Schematic diagram of the internal structure of a steel intelligent classification system using confidence comparison according to Embodiment C of the present invention.

[0040] and Figure 1 different, Figure 3 The steel intelligent classification system based on confidence comparison can also include the following components:

[0041] A power supply device, connected to the directional analysis mechanism, the affine transformation component, the filter execution component and the data sharpening component, respectively, for respectively performing real-time power supply of the respective working voltages of the directional analysis mechanism, the affine transformation component, the filter execution component and the data sharpening component;

[0042] Wherein, the power supply device is respectively connected to the directional analysis mechanism, the affine transformation component, the filter execution component and the data sharpening component, and is used to respectively perform real-time power supply of the respective working voltages of the directional analysis mechanism, the affine transformation component, the filter execution component and the data sharpening component, including: the power supply device is an uninterruptible power supply device;

[0043] And wherein, the power supply device is respectively connected to the directional analysis mechanism, the affine transformation component, the filtering execution component and the data sharpening component, and is used to respectively perform real-time power supply for the respective working voltages of the directional analysis mechanism, the affine transformation component, the filtering execution component and the data sharpening component, including: any working voltage of the directional analysis mechanism, the affine transformation component, the filtering execution component and the data sharpening component is 3.3V or 5V.

[0044] In addition, in the intelligent steel classification system using confidence comparison, a trained feedforward neural network is used to intelligently identify the confidence of the existence of the pattern to be compared in the layer-by-layer conversion image based on each grayscale value corresponding to each pixel point of the layer-by-layer conversion image, each coordinate data corresponding to each pixel point of the layer-by-layer conversion image, each grayscale value corresponding to each pixel point of the pattern to be compared, and each coordinate data corresponding to each pixel point of the pattern to be compared, which also includes: executing the trained feedforward neural network to obtain the confidence of the existence of the pattern to be compared in the layer-by-layer conversion image output by the trained feedforward neural network.

[0045] Although the above description may be considered as the best mode and / or other examples, it should be understood that various modifications may be made herein and the subject matter disclosed herein may be implemented in various forms and examples, and the teachings may be applied in numerous applications, of which only some are disclosed herein. The following claims are intended to claim any application, modification, and variation that falls within the true scope of the present teachings.

Claims

1. An intelligent steel classification system using confidence comparison, characterized in that: The system comprises: Classification and transmission equipment, used for transmitting the target steel to be classified to the top of the visual collection station, wherein the target steel to be classified is one of round steel, square steel, I-beam, unequal angle steel and hexagonal steel; A network sensing device connected to the classification transmission device is used to perform a photoelectric sensing action on the cross section of the target steel material to be classified based on a network control instruction after detecting that the classification transmission device has completed transmission, so as to obtain and output a corresponding real-time sensing image; A layer-by-layer conversion device, connected to the network sensing device and comprising an affine transformation component, a filtering execution component and a data sharpening component, for continuously performing affine transformation processing, salt and pepper noise removal processing and spatial differential method sharpening processing on the received real-time sensing image to obtain and output a corresponding layer-by-layer conversion image; A directional parsing mechanism and the layer-by-layer conversion device are used to receive the layer-by-layer conversion image, parse the grayscale values ​​corresponding to the pixels of the layer-by-layer conversion image and the coordinate data corresponding to the pixels of the layer-by-layer conversion image, obtain the grayscale values ​​corresponding to the pixels of the pattern to be compared and the coordinate data corresponding to the pixels of the pattern to be compared, and use a trained feedforward neural network to intelligently identify the confidence of the pattern to be compared in the layer-by-layer conversion image based on the grayscale values ​​corresponding to the pixels of the layer-by-layer conversion image, the coordinate data corresponding to the pixels of the layer-by-layer conversion image, the grayscale values ​​corresponding to the pixels of the pattern to be compared and the coordinate data corresponding to the pixels of the pattern to be compared; Wherein, the confidence of the pattern to be compared in the layer-by-layer conversion image is intelligently identified by using a trained feedforward neural network based on each grayscale value corresponding to each pixel point of the layer-by-layer conversion image, each coordinate data corresponding to each pixel point of the layer-by-layer conversion image, each grayscale value corresponding to each pixel point of the pattern to be compared, and each coordinate data corresponding to each pixel point of the pattern to be compared, including: the pattern to be compared is one of a circle, a square, an I-shaped, an unequal-sided triangle and a hexagon, and the patterns to be compared are respectively subjected to corresponding intelligent identification processing by applying the trained feedforward neural network to obtain each confidence, and the steel type corresponding to the pattern to be compared with the highest confidence is used as the current classification type of the target steel to be classified; Among them, using a trained feedforward neural network to intelligently identify the confidence of the existence of the pattern to be compared in the layer-by-layer converted image based on each grayscale value corresponding to each pixel point of the layer-by-layer converted image, each coordinate data corresponding to each pixel point of the layer-by-layer converted image, each grayscale value corresponding to each pixel point of the pattern to be compared, and each coordinate data corresponding to each pixel point of the pattern to be compared also includes: the number of times the feedforward neural network is trained is positively correlated with the resolution of the layer-by-layer converted image.

2. The intelligent steel classification system using confidence comparison as claimed in claim 1, characterized in that: Using a trained feedforward neural network to intelligently identify the confidence of the existence of the pattern to be compared in the layer-by-layer converted image based on the grayscale values ​​corresponding to the pixels of the layer-by-layer converted image, the coordinate data corresponding to the pixels of the layer-by-layer converted image, the grayscale values ​​corresponding to the pixels of the pattern to be compared, and the coordinate data corresponding to the pixels of the pattern to be compared also includes: simultaneously inputting the grayscale values ​​corresponding to the pixels of the layer-by-layer converted image, the coordinate data corresponding to the pixels of the layer-by-layer converted image, the grayscale values ​​corresponding to the pixels of the pattern to be compared, and the coordinate data corresponding to the pixels of the pattern to be compared into the trained feedforward neural network.

3. The steel intelligent classification system using confidence comparison as claimed in claim 2, characterized in that: The system further comprises: a content enhancement mechanism, connected to the directional parsing mechanism, the affine transformation component, the filter execution component and the data sharpening component, respectively, for performing content enhancement on the respective output image signals of the directional parsing mechanism, the affine transformation component, the filter execution component and the data sharpening component, so as to obtain respective content enhanced images corresponding to the respective output image signals of the directional parsing mechanism, the affine transformation component, the filter execution component and the data sharpening component; Among them, content enhancement is performed on the respective output image signals of the directional parsing mechanism, the affine transformation component, the filtering execution component and the data sharpening component to obtain respective content enhanced images corresponding to the respective output image signals of the directional parsing mechanism, the affine transformation component, the filtering execution component and the data sharpening component, including: the first level content enhancement performs geometric correction actions on the respective output image signals of the directional parsing mechanism, the affine transformation component, the filtering execution component and the data sharpening component.

4. The intelligent steel classification system using confidence comparison as claimed in claim 3, characterized in that: The content enhancement is performed on the respective output image signals of the directional parsing mechanism, the affine transformation component, the filtering execution component and the data sharpening component to obtain respective content enhanced images corresponding to the respective output image signals of the directional parsing mechanism, the affine transformation component, the filtering execution component and the data sharpening component. The second level content enhancement also performs morphological processing actions on the respective output image signals of the directional parsing mechanism, the affine transformation component, the filtering execution component and the data sharpening component.

5. The intelligent steel classification system using confidence comparison as claimed in claim 4, characterized in that: The content enhancement is performed on the respective output image signals of the directional parsing mechanism, the affine transformation component, the filtering execution component and the data sharpening component to obtain respective content enhanced images corresponding to the respective output image signals of the directional parsing mechanism, the affine transformation component, the filtering execution component and the data sharpening component. The third level content enhancement also includes adaptive recursive filtering actions on the respective output image signals of the directional parsing mechanism, the affine transformation component, the filtering execution component and the data sharpening component.

6. The steel intelligent classification system using confidence comparison as claimed in claim 5, characterized in that: The content enhancement is performed on the respective output image signals of the directional parsing mechanism, the affine transformation component, the filtering execution component and the data sharpening component to obtain respective content enhanced images corresponding to the respective output image signals of the directional parsing mechanism, the affine transformation component, the filtering execution component and the data sharpening component. The fourth level content enhancement also performs image enhancement actions based on exponential transformation on the respective output image signals of the directional parsing mechanism, the affine transformation component, the filtering execution component and the data sharpening component.

7. The intelligent steel classification system using confidence comparison as described in any one of claims 3 to 6, characterized in that: The system further comprises: The power supply device is respectively connected to the directional analysis mechanism, the affine transformation component, the filtering execution component and the data sharpening component, and is used to respectively perform real-time power supply for the respective working voltages of the directional analysis mechanism, the affine transformation component, the filtering execution component and the data sharpening component.

8. The intelligent steel classification system using confidence comparison as claimed in claim 7, characterized in that: A power supply device is respectively connected to the directional analysis mechanism, the affine transformation component, the filtering execution component and the data sharpening component, and is used to respectively perform real-time power supply for the respective working voltages of the directional analysis mechanism, the affine transformation component, the filtering execution component and the data sharpening component, including: the power supply device is an uninterruptible power supply device.

9. The intelligent steel classification system using confidence comparison as claimed in claim 7, characterized in that: The power supply device is respectively connected to the directional resolution mechanism, the affine transformation component, the filtering execution component and the data sharpening component, and is used to respectively perform real-time power supply for the respective working voltages of the directional resolution mechanism, the affine transformation component, the filtering execution component and the data sharpening component, including: any working voltage of the directional resolution mechanism, the affine transformation component, the filtering execution component and the data sharpening component is 3.3V or 5V.

Images