Intelligent identification system and method for building material categories

Abstract

The invention discloses an intelligent recognition system and method for building material categories, and belongs to the technical field of building material recognized.The system comprises a combined light emitting module, an image collecting module, a data processing module and a recognition result output module, the combined light emitting module emits combined light containing light of various different wavelengths, and the image collecting module is used for collecting the combined light; the image acquisition module acquires image information of a building material under combined light irradiation, the data processing module processes and analyzes an image based on an electric digital data processing technology, the identification result output module outputs an identification result, and the method comprises the steps of combined light irradiation, image acquisition, data processing, mode identification and result output. According to the invention, the identification degree of building material characteristics is improved by using combined light, intelligent identification is realized by combining an electric digital data processing technology, the identification efficiency and accuracy are improved, and the use requirements of different scenes are met.

Description

Technical Field

[0001] The present invention relates to the technical field of building material category identification, and in particular to an intelligent identification system and method for building material category identification. Background Art

[0002] In the construction industry, there are many types of building materials. Accurately identifying the categories of building materials is of great significance for the quality control, material management, and subsequent maintenance and renovation of construction projects. Traditional building material identification methods mainly rely on manual observation and experience judgment. This method is inefficient and easily affected by subjective factors, resulting in low accuracy and reliability of identification results. With the development of science and technology, recognition technology based on a single light source has emerged, which uses visible light or infrared light for identification. Since the characteristic differences of different building materials under a single light source are not obvious enough, the recognition effect is not ideal. Therefore, an intelligent recognition system and method that can improve the efficiency and accuracy of building material recognition is needed.

[0003] The purpose of the present invention is to provide an intelligent identification system and method for building material categories to solve the problems of low efficiency, poor accuracy and unsatisfactory single light source identification effect of the traditional identification method proposed in the above background technology. Summary of the Invention

[0004] To achieve the above objectives, the present invention proposes an intelligent identification system and method for building material categories, comprising: Combined light emission module: emits light of multiple wavelengths to the building material to be identified, and obtains the reflection information of the building material under different optical properties; Image acquisition module: The image acquisition module is set on one side of the combined light emitting module, collects image information of building materials under the combined light irradiation, and converts the collected image information into electrical signals; Preprocessing module: After the image acquisition module obtains the image data, the preprocessing module performs denoising, enhancement, and correction on the image data, and then transmits the processed image data to the data fusion module; Data fusion module: fuses multiple sets of data with different optical characteristics obtained by the two modules to extract comprehensive feature information; Feature comparison module: A feature database of various building materials is pre-stored. The comprehensive feature information extracted by the data fusion module is compared and analyzed with the features in the database to determine the category of the building material to be identified; Result output module: outputs the category information of building materials based on the analysis results of the feature comparison module.

[0005] In one example, the combined light emission module includes multiple light sources with different parameters, which are evenly distributed around the building material to be identified, and the emission angle is adjustable to fully cover the surface of the building material and obtain rich optical reflection information.

[0006] In one example, the image acquisition module uses a high-resolution, high-sensitivity image sensor, which can clearly capture reflected light images of building materials under different lighting conditions, and has automatic gain control and white balance adjustment functions to ensure the stability of image quality. The data fusion module uses a multimodal fusion algorithm to jointly model the wavelength, polarization state, and intensity optical parameter information of the combined light with the spatial features and texture features of the image to generate a comprehensive feature vector containing optical characteristics and visual features.

[0007] In one example, the feature database in the feature comparison module is regularly updated and maintained, and new building material sample data is continuously collected and feature extracted and annotated, and added to the database to improve the system's ability to identify new building materials.

[0008] In one example, the result output module includes a display screen and a data transmission interface. The display screen can intuitively display the category information of the building materials, and the data transmission interface can transmit the recognition results to an external device to facilitate subsequent data processing and application.

[0009] In one example, the preprocessing module uses a median filtering algorithm to perform denoising, uses a histogram equalization algorithm to perform image enhancement, and uses a geometric correction algorithm to correct the image to eliminate the influence of shooting angle and distance factors on the image.

[0010] A method for intelligently identifying building material categories, comprising the following steps: Combined light emission and reflection information acquisition: emitting light of multiple wavelengths to the building material to be identified, and obtaining reflection information of the building material under different optical properties; Image acquisition and conversion: The image acquisition module acquires image information of building materials under combined light illumination; converts the acquired image information into electrical signals, including visual feature information of the building materials’ color, texture, and shape; Image preprocessing: After the image acquisition module acquires data, the preprocessing module is started. The preprocessing module performs denoising on the image data, then performs enhancement after denoising, and then performs correction after enhancement. The processed image data is then transmitted to the data fusion module. Data fusion processing: Receive the reflection information data under different optical characteristics obtained by the combined light emission module and the image data output by the pre-processing module, fuse these two sets of data under different optical characteristics, and extract comprehensive feature information; Feature comparison analysis: Compare and analyze the comprehensive feature information extracted by the data fusion module with the features in the database, and determine the degree of consistency between the characteristics of the building material to be identified and the various building materials in the database by calculating the similarity and matching indicators; Result output: Based on the analysis results of the feature comparison module, the result output module is started. The result output module outputs the category information of the building materials, clearly displays the specific category of the building materials to be identified, and completes the intelligent identification process of the building material category.

[0011] In one example, in the combined light irradiation step, the reflection information of the combined light emitting module includes reflection intensity and reflection spectrum characteristic data of the building material to light under irradiation of light of different wavelengths.

[0012] In one example, in the data processing step, the characteristic parameters include color, texture, and spectral reflectance.

[0013] The intelligent identification system and method for building material categories proposed by the present invention can bring the following beneficial effects: 1. The present invention uses a combined light emitting module to emit a combination of light containing multiple different wavelengths. This can fully utilize the differences in the optical properties of building materials under different wavelengths of light, improve the recognition of building material features in the image, and thus improve the accuracy and reliability of the recognition results.

[0014] 2. The present invention uses a data processing module based on electronic digital data processing technology to efficiently and accurately process and analyze images, realize intelligent recognition of building material categories, greatly improve recognition efficiency, and reduce manual intervention.

[0015] 3. The present invention provides multiple output modes through the recognition result output module, which makes it convenient for users to obtain recognition results and meets the usage requirements in different scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings: Figure 1 The figure is a schematic structural diagram of an intelligent identification system for building material categories according to the present invention.

[0017] Figure 2 The figure is a flow chart of an intelligent identification method for building material categories according to the present invention. DETAILED DESCRIPTION

[0018] In order to more clearly and completely illustrate the technical solution of the present invention, the present invention will be further described below with reference to the accompanying drawings.

[0019] Please refer to Figure 1 and Figure 2 The present invention relates to an intelligent identification system and method for building material categories, which combines light emission, image acquisition, preprocessing, data fusion, feature comparison and result output modules.

[0020] The combined light emission module contains multiple light sources with different parameters, which are evenly distributed around the building materials to be identified. LED lights with different wavelength ranges are selected as light sources, covering multiple bands from ultraviolet to infrared. The ultraviolet light source is used to detect certain special fluorescent properties on the surface of building materials, the visible light source reflects the general color and appearance characteristics of the material, and the infrared light source helps to detect the internal structure and thermal properties of the material.

[0021] In practical applications, the emission angle of the light source is adjusted according to the size and shape of the building materials so that the light is evenly irradiated on all parts of the material, thereby obtaining rich optical reflection information. The combined light emission module is equipped with a light intensity control system, which automatically adjusts the brightness of the light source according to changes in ambient light. Under different ambient lighting conditions, the light intensity emitted to the building materials remains consistent. In a strong light environment, the brightness of the light source is automatically reduced, and in a weak light environment, the brightness of the light source is appropriately increased.

[0022] The image acquisition module uses a high-resolution, high-sensitivity image sensor, and uses a CMOS image sensor with tens of millions of pixels, which can clearly capture the subtle details of building materials under combined light. This sensor is highly adaptable to different lighting conditions and can ensure image clarity and accuracy under different light intensities.

[0023] The image processing functions include automatic gain control and white balance adjustment. The automatic gain control can automatically adjust the image magnification according to the brightness of the environment to ensure that the image brightness is moderate under different lighting conditions. The white balance adjustment function can eliminate the influence of ambient light on the image color, making the collected image color true and accurate. In an environment with mixed natural light and artificial light, through white balance adjustment, the color of the building materials in the image is consistent with the actual color, which is convenient for subsequent feature extraction and analysis.

[0024] When the combined light shines on the building materials, the image acquisition module starts working to collect the reflected light image of the building materials. During the acquisition process, the sensor converts the light signal into an electrical signal and performs preliminary amplification and filtering on the electrical signal to improve the image quality. The collected image information contains the color, texture, and shape visual feature information of the building materials, which will serve as an important basis for subsequent identification.

[0025] The preprocessing module first denoises the image data using the median filtering algorithm. Median filtering is a nonlinear filtering method that can effectively remove salt and pepper noise and speckle noise in images. The specific implementation process is that for each pixel in the image, the median value of the grayscale values ​​of the pixels in a certain neighborhood around it is taken as the new grayscale value of the pixel. For a 3×3 neighborhood, the grayscale values ​​of the 9 pixels are sorted from small to large, and the 5th value is taken as the new value of the center pixel. In this way, noise interference can be removed while retaining the edge and detail information of the image.

[0026] After denoising, image enhancement processing is performed using the histogram equalization algorithm. Histogram equalization enhances the contrast of the image by adjusting the grayscale distribution of the image. It stretches the grayscale histogram of the original image to make the grayscale distribution of pixels in the image more uniform, thereby improving the overall clarity and recognizability of the image. For an image of building materials with grayscale concentrated in darker areas, after histogram equalization, the details of both the bright and dark parts of the image can be better displayed, which is conducive to the subsequent feature extraction.

[0027] Finally, image correction processing is performed using a geometric correction algorithm. Since the shooting angle and distance may cause the image to deform and distort during the actual shooting process, the geometric correction algorithm corrects the image by establishing a mapping relationship between the image coordinates and the actual space coordinates. For image deformation caused by the tilt of the shooting angle, the image is rotated and scaled by calculating the transformation matrix to restore it to the correct perspective and eliminate the influence of the shooting factors on the image. The processed image data will be transmitted to the data fusion module to provide an accurate data basis for subsequent comprehensive feature extraction.

[0028] The data fusion module uses a multimodal fusion algorithm to fuse the data under different optical characteristics obtained by the combined light emission module with the image data output by the image acquisition module. Specifically, the wavelength, polarization state, and intensity optical parameter information of the combined light are jointly modeled with the spatial characteristics and texture characteristics of the image. For example, the reflection intensity data of building materials under light of different wavelengths is combined with the texture characteristics of the corresponding area in the image to generate a comprehensive feature vector that includes optical characteristics and visual characteristics.

[0029] Feature extraction and fusion: feature extraction is performed on data of different modalities. For optical parameter information, its reflection intensity characteristics and spectral curve characteristics at different wavelengths are extracted; for image data, its texture characteristics and shape characteristics are extracted. Then, these features are fused according to certain weights and rules to form a comprehensive feature representation. This comprehensive feature vector can more comprehensively and accurately describe the characteristics of building materials, providing richer and more discriminative information for subsequent feature comparison.

[0030] The feature comparison module pre-stores a feature database of various building materials. This database is constructed by collecting a large number of building material samples of known categories and extracting feature data after the data acquisition, preprocessing, and fusion steps of the above system. Each sample in the database is labeled with the building material category to which it belongs, such as cement, steel, wood, and glass.

[0031] After the data fusion module outputs the comprehensive feature information of the building material to be identified, the feature comparison module compares and analyzes it with the features in the database. During the comparison and analysis process, the method of calculating similarity and matching indicators is adopted to determine the degree of consistency between the features of the building material to be identified and the features of various building materials in the database by comparing the sizes of these indicators. If the similarity between the features to be identified and the features of a certain type of building material is the highest and exceeds the set threshold, the building material to be identified is considered to belong to that category.

[0032] Database updating and maintenance: by continuously collecting new building material sample data, extracting and labeling them according to the same process, adding them to the database, cleaning and optimizing the existing data in the database, deleting outdated or inaccurate data, and ensuring the quality and validity of the database.

[0033] The result output module includes a display screen, which can intuitively display the category information of building materials. The display screen can use an LCD screen or a touch screen device with clear display effects and good interactivity. When the system completes the recognition, the category name and related characteristic parameter information of the building materials are displayed in the form of text, icons or pictures on the display screen, making it convenient for users to directly view the recognition results.

[0034] The result output module is also equipped with a data transmission interface that can transmit the recognition results to an external device. The data transmission interface can be a USB interface, a network interface or a serial port. Through these interfaces, the recognition results can be transmitted to a computer, a mobile device or other data processing system to facilitate subsequent data processing, storage and application. At the construction site, the recognition results can be transmitted to the remote monitoring center through the network interface to realize real-time management and monitoring of the use of construction materials.

[0035] The specific implementation steps of the intelligent recognition method are combined light irradiation and data acquisition, image acquisition and conversion, preprocessing process, data fusion and feature extraction, feature comparison and analysis, and result output and display.

[0036] Emitting multiple wavelengths of light, starting the combined light emission module, and emitting multiple wavelengths of light to the building materials to be identified. These lights cover different bands from ultraviolet to infrared, and can stimulate the reflection of building materials under different optical properties. For example, when detecting wood, ultraviolet light can reveal the distribution of fluorescent substances on the surface of the wood, visible light shows the color and texture of the wood, and infrared light reflects the internal fiber structure of the wood.

[0037] To obtain reflection information, the combined light transmitting module receives light reflected by building materials while emitting light, and obtains reflection information. The reflection information includes the reflection intensity of building materials to light under different wavelengths of light and the reflection spectrum characteristic data. For example, for a certain metal building material, it may have a higher reflection intensity at a specific wavelength, and its reflection spectrum has a unique waveform. This information will be recorded as an important basis for subsequent identification.

[0038] Collect image information. The image acquisition module starts working under the combined light to collect image information of building materials. The high-resolution and high-sensitivity sensors of the image acquisition module can clearly capture the color, texture, and shape visual feature information of building materials. For example, when collecting tile images, it can accurately record the pattern on the tile surface, color depth changes, and edge texture details.

[0039] The collected image information is converted into electrical signals by the image sensor. This process is achieved through the principle of photoelectric conversion, that is, the image sensor converts light intensity information into corresponding electrical signal intensity. The converted electrical signal contains complete information about the building material image, providing a data basis for subsequent processing.

[0040] The preprocessing module first denoises the image data by using the median filtering algorithm to process each pixel in the image and its surrounding neighborhood. For example, for a concrete image interfered with by noise, the median filter can effectively remove the random noise points in the image, making the image clearer and smoother while retaining the real texture and structural information of the concrete surface.

[0041] After denoising, image enhancement processing is performed using a histogram equalization algorithm, which transforms the grayscale histogram of the image to make the grayscale distribution of the image more uniform. For example, for an image of building materials in a shadow area, histogram equalization can enhance the contrast of the image, making the details of the shadow part clearer, improving the overall quality of the image, and facilitating subsequent feature extraction and analysis.

[0042] Finally, image correction processing is performed using a geometric correction algorithm. Since the shooting angle and distance may cause image deformation, the geometric correction algorithm corrects the image by establishing a mapping relationship between the image coordinates and the actual space coordinates. For example, for images where the shape of building materials is distorted due to the tilted shooting angle, after geometric correction, their true shape and proportion can be restored, eliminating the influence of shooting factors on the image and ensuring the accuracy of subsequent data processing.

[0043] Data reception and fusion preparation: The data fusion module receives the reflection information data under different optical characteristics obtained by the combined light emission module and the image data output by the preprocessing module. After receiving the data, it organizes and preprocesses the data to ensure that the data format and accuracy meet the fusion requirements, normalizes the reflection intensity data, and converts the image data into a unified color space and resolution.

[0044] Multimodal fusion and feature extraction uses a multimodal fusion algorithm to jointly model the wavelength, polarization state, and intensity optical parameter information of the combined light with the spatial features and texture features of the image. During the fusion process, feature extraction is performed on data of different modalities respectively. For optical parameter information, its reflection intensity characteristics and spectral curve characteristics at different wavelengths are extracted; for image data, its texture features, grayscale co-occurrence matrix features, shape features, edge contours, and geometric shape parameters are extracted. Then, these features are fused according to certain weights and rules to form a comprehensive feature vector. For example, when identifying glass materials, the high transmittance characteristics of glass at different wavelengths are combined with the smooth surface texture characteristics of the glass in the image to generate a comprehensive feature vector that can accurately describe the characteristics of the glass.

[0045] Database query and comparison: Compare and analyze the comprehensive feature information extracted by the data fusion module with the features in the database. The feature comparison module queries relevant feature data from the pre-stored database of various building material features and compares them one by one with the features to be identified. For example, for the insulation material to be identified, its comprehensive features are compared with the features of various insulation materials, polystyrene foam, and rock wool in the database, and the similarity index between them is calculated.

[0046] Similarity calculation and judgment. During the comparative analysis process, methods for calculating similarity and matching indicators are adopted. Commonly used similarity calculation methods include Euclidean distance and cosine similarity. By calculating the similarity index between the feature to be identified and the features of each category in the database, the degree of consistency between the building material to be identified and the features of each category of building materials is judged. For example, if the cosine similarity between the feature to be identified and the feature of a certain type of insulation material is the highest and exceeds the set threshold, then the building material to be identified is considered to belong to this type of insulation material.

[0047] Result output and display: Start the result output module, and according to the analysis results of the feature comparison module, start the result output module. If the feature comparison results clearly indicate that the building material to be identified belongs to a certain category and the similarity exceeds the set threshold, the result output module is ready to output the corresponding category information.

[0048] Display category information and data transmission. The display screen of the result output module intuitively displays the category information of the building materials. For example, the words "thermal insulation material polystyrene foam" are displayed on the screen, and the relevant characteristic parameters of the material, such as thermal conductivity and density, can be displayed. In addition, the recognition results are transmitted to external devices through the data transmission interface. For example, the recognition results are transmitted to a computer system for further data analysis and recording, or transmitted to a mobile device, so that on-site staff can view and use the recognition results at any time to complete the intelligent recognition process of the building material category.

[0049] In this embodiment, the combined light emission module includes multiple light sources with different parameters, including LEDs, halogen lamps, and infrared light sources. The wavelengths cover the visible light range of 400-760nm and the near-infrared range of 760-1100nm. The light sources are evenly distributed around the material to be identified in a ring-shaped or hemispherical array. The emission angle can be dynamically adjusted within the range of 0-90° through an electric adjustment mechanism to ensure full-angle illumination of the material surface and obtain multi-dimensional optical reflection information including reflection intensity, spectral characteristics, and polarization state. The image acquisition module uses a CMOS image sensor with high resolution ≥12 million pixels, high sensitivity, and quantum efficiency ≥80%. It integrates automatic gain control and adaptive white balance modules, can stably capture images within the light intensity range of 100lux-5000lux, and outputs raw data including RGB three channels and near-infrared channel, with a resolution of 4096×3072 or above. The preprocessing module performs the following steps in sequence: median filtering denoising: using a 3×3 or 5×5 sliding window to remove salt and pepper noise, retain edge details, and perform global histogram equalization on the grayscale image to enhance contrast and highlight texture features. The camera's intrinsic and extrinsic parameters are used to correct perspective distortion and eliminate the geometric effects of shooting distance, 0.5-2 meters, and ±30° tilt on the image. The multispectral reflectance intensity data were aligned with the image pixel values ​​to generate a multi-channel data cube, and the LBP texture features and Gabor directional features of the image were extracted. Combined with the spectral reflectance curve of 200-1100 nm with an interval of 10 nm, the dimension was reduced to 50 dimensions through principal component analysis. The recognition results under different light source conditions are fused with confidence to generate a 100-dimensional comprehensive feature vector containing optical properties, wavelength, polarization, intensity and visual features, color, texture and shape.

[0050] The feature comparison module has a built-in dynamically updateable feature database stored in XML or JSON format, including: Basic characteristics: material type, cement, steel, stone, glass, model, specification; Optical characteristics: reflectivity mean, variance, and peak wavelength at each wavelength; Visual features: color histogram, texture energy, shape invariant moment Hu moment, and the comparison algorithm uses improved cosine similarity.

[0051] The result output module integrates a 7-inch high-definition LCD touch screen, which displays material category, confidence level, percentage, and feature matching curve in real time. The data transmission interface includes USB3.0, RJ45 Ethernet and WiFi modules, which support real-time transmission of recognition results in JSON format to the PLC control system or cloud database. The interface protocol complies with the OPCUA standard. A method for intelligently identifying building material categories, the method comprising the following steps: Multi-spectral illumination and reflection information collection: the combined light emission module sequentially emits polarized light of 6 wavelengths: 450nm, 550nm, 650nm, 850nm, 950nm, and 1050nm. Linear polarization / circular polarization is optional. The irradiation time of each light source is 50ms. The intensity of the reflected light on the surface of the material is synchronously collected through the photoelectric detector to generate a reflection spectrum matrix containing 120 spectral data points from 200-1100nm and an interval of 10nm. Image acquisition and electrical signal conversion: The image sensor performs multi-frame exposure under combined light illumination. Short exposure: 1 / 1000s is used for strong light scenes, and long exposure: 1 / 30s is used for low light scenes. It is converted into a 14-bit digital signal through the on-chip ADC and outputs raw image data containing four channels of R, G, B, and NIR. The shooting time and light source parameter metadata are simultaneously recorded.

[0052] For image preprocessing, median filtering is performed on the NIR channel. The window size is dynamically adjusted according to the noise density. A 5×5 window is used when the noise rate is greater than 5%. Histogram equalization is performed on the RGB channels and the image is processed in blocks of 8×8 pixels to preserve local contrast. Perspective correction is performed on the deformed image based on the pre-calibrated camera parameters. Bilinear interpolation is used for resampling to ensure that the resolution error of the corrected image is less than 0.1%. Multimodal data fusion is performed to convert the preprocessed image into the HSV color space, extract the 180-bin histogram of the H channel and the texture energy of the S channel, perform Savitzky-Golay smoothing on the spectral data, calculate the first-order derivative of the reflectance of each band, enhance the difference in spectral features, and use the feature-level fusion algorithm to splice the visual features and spectral features and input them into the autoencoder to compress and generate an 80-dimensional comprehensive feature vector. Feature comparison and category determination: retrieve feature clusters of similar materials in the feature database, use the K-nearest neighbor algorithm with K=3 to calculate the weighted distance between the feature to be identified and the features in the library, set the confidence threshold, the default is 0.85, and when the highest similarity ≥ the threshold, output the matching category; otherwise, mark it as "unknown material", triggering the data collection process to update the database.

[0053] For result output and application, the display interface presents material categories in a tree structure ("metal materials → steel materials → hot-rolled steel plates"), and highlights the three candidate categories with the highest matching degree. The data interface supports timed, 1-second / time, or triggered transmission. The transmission content includes recognition results, original image path, and acquisition timestamp, facilitating subsequent quality traceability and production process control.

[0054] Of course, the present invention may have many other implementations. Based on this implementation, other implementations obtained by ordinary technicians in this field without any creative work are all within the scope of protection of the present invention.

Claims

1. An intelligent identification system for building material categories, characterized in that: include: Combined light emission module: emits light of multiple wavelengths to building materials to obtain reflection information of building materials under different optical properties; Image acquisition module: The image acquisition module is set on one side of the combined light emitting module, collects image information of building materials under the combined light irradiation, and converts the collected image information into electrical signals; Preprocessing module: After the image acquisition module obtains the image data, the preprocessing module performs denoising, enhancement, and correction on the image data, and then transmits the processed image data to the data fusion module; Data fusion module: fuses multiple sets of data with different optical characteristics obtained by the two modules to extract comprehensive feature information; Feature comparison module: A feature database of various building materials is pre-stored. The comprehensive feature information extracted by the data fusion module is compared and analyzed with the features in the database to determine the category of the building material to be identified; Result output module: outputs the category information of building materials based on the analysis results of the feature comparison module.

2. The intelligent identification system for building material categories according to claim 1, characterized in that: The combined light emission module includes multiple light sources with different parameters. The light sources are evenly distributed around the building material to be identified. The emission angle is adjustable, fully covering the surface of the building material and obtaining rich optical reflection information.

3. The intelligent identification system for building material categories according to claim 1, characterized in that: The image acquisition module uses a high-resolution, high-sensitivity image sensor to clearly capture reflected light images of building materials under different lighting conditions. It has automatic gain control and white balance adjustment functions to ensure the stability of image quality. The data fusion module uses a multimodal fusion algorithm to jointly model the wavelength, polarization state, and intensity optical parameter information of the combined light with the spatial characteristics and texture characteristics of the image to generate a comprehensive feature vector containing optical characteristics and visual characteristics.

4. The intelligent identification system for building material categories according to claim 1, characterized in that: The feature database in the feature comparison module is regularly updated and maintained. By continuously collecting new building material sample data and performing feature extraction and annotation, it is added to the database to improve the system's ability to identify new building materials.

5. The intelligent identification system for building material categories according to claim 1, characterized in that: The result output module includes a display screen and a data transmission interface. The display screen can intuitively display the category information of the building materials, and the data transmission interface can transmit the recognition results to an external device to facilitate subsequent data processing and application.

6. The intelligent identification system for building material categories according to claim 1, characterized in that: The pre-processing module adopts a median filtering algorithm to perform denoising, adopts a histogram equalization algorithm to perform image enhancement, and adopts a geometric correction algorithm to correct the image, thereby eliminating the influence of shooting angle and distance factors on the image.

7. An intelligent identification method for building material categories, characterized in that: The following steps are involved: Step 1: Emit light of multiple wavelengths to the building material to be identified to obtain reflection information of the building material under different optical properties; Step 2: The image acquisition module acquires image information of the building materials under the combined light illumination and converts the acquired image information into electrical signals, including visual feature information of the color, texture, and shape of the building materials; Step 3: After the image acquisition module acquires the data, the pre-processing module is started. The pre-processing module performs denoising on the image data, then performs enhancement after denoising, and then performs correction after enhancement. The processed image data is then transmitted to the data fusion module. Step 4: Receive the reflection information data under different optical characteristics obtained by the combined light emitting module and the image data output by the pre-processing module, fuse these two or more sets of data under different optical characteristics, and extract comprehensive feature information; Step 5: Compare and analyze the comprehensive feature information extracted by the data fusion module with the features in the database, and determine the degree of consistency between the characteristics of the building material to be identified and the various building materials in the database by calculating the similarity and matching indexes; Step 6: Based on the analysis results of the feature comparison module, the result output module is started. The result output module outputs the category information of the building materials, clearly displays the specific category of the building materials to be identified, and completes the intelligent identification process of the building material category.

8. The intelligent identification method for building material categories according to claim 7, characterized in that: In the combined light irradiation step, the reflection information of the combined light emitting module includes the reflection intensity and reflection spectrum characteristic data of the building material to light under the irradiation of light of different wavelengths.

9. The intelligent identification method for building material categories according to claim 7, characterized in that: In the data processing step, the characteristic parameters include color, texture, and spectral reflectance.

Images