A dental caries detection device based on photometry and visual analysis

Abstract

A dental caries detection device based on photometry and visual analysis includes three rotating platforms, which are equivalent to three modules, namely a structured light scanning module, a visual recognition module, and a light sectioning roughness detection module; the structured light scanning module obtains a high-precision three-dimensional structural model of each tooth surface; the visual recognition module obtains high-precision image color data of each tooth surface and identifies slight color changes on the tooth surface; the light sectioning detection module emits a vertical light band through a line laser to illuminate the target tooth surface, and the reflected light passes through the lens and is imaged on the surface of a microscope camera to obtain the tooth surface roughness distribution; the present invention can establish a tooth surface contour and parameter model through multi-source data collaborative processing, and can analyze and diagnose the multi-source data model obtained by scanning by combining big data machine learning and artificial intelligence analysis, determine whether the scanned tooth surface has caries, and locate the caries position.

Description

Technical Field

[0001] The present invention relates to the technical field of caries detection, and in particular to a caries detection device based on photometry and visual analysis for detecting maxillofacial caries with high precision and low cost. Background Art

[0002] A caries detector is a device used in dental clinics to detect and determine whether a patient has caries. Commonly used caries detectors in dental clinics include traditional imaging equipment such as oral and maxillofacial X-ray machines, medical CT machines, and oral and maxillofacial cone-beam CT machines. However, due to my country's large population base and high caries prevalence, higher requirements are placed on the accessibility and practicality of caries detection devices. For example, residents' oral health self-examinations require miniaturized caries detectors; non-professionals performing caries detection require caries detectors with intelligent caries diagnosis capabilities; and special populations require caries detectors that operate in a radiation-free mode.

[0003] Traditional caries detection instruments can achieve X-ray imaging and three-dimensional reconstruction of patients' teeth through radiation and scanning, and can intuitively and accurately display the internal characteristics and diseases of teeth. However, due to their working method, there are still many problems: 1. Traditional caries detection instruments mainly diagnose by detecting the geometric characteristics of teeth, but because there are only local color changes on the jaw surface in the early stages of caries, traditional detection instruments are difficult to make effective diagnoses, causing patients to miss the best treatment time; 2. Traditional caries detection instruments are large and expensive, and the test results need to be analyzed by professionals, and cannot be used by home residents for self-examination. With the significant increase in the number of patients and the growth in residents' demand for daily oral health examinations, traditional caries detection methods are gradually unable to meet existing detection needs; 3. Traditional medical imaging examinations use X-ray imaging, which carries the risk of ionizing radiation. Although the radiation dose is low, it still causes concern for some patients, especially children and pregnant women.

[0004] To address this, numerous scholars at home and abroad have proposed the following solutions: using terahertz scanning imaging systems to scan decayed teeth and image the carious tissue; using optical coherence tomography to accurately identify diseased tissue; and using fluorescence analysis to image caries and visualize the extent of tooth decay. While these methods offer advantages such as high resolution, lack of radiation, and visualization, they still face the following challenges: 1. The equipment is bulky and expensive, significantly limiting its application and widespread adoption; 2. The use of these testing devices requires a high level of expertise, and the feedback is often presented in specialized tables or data stacks that require analysis by professionals, making them inadequate for the daily oral health needs of ordinary residents. Summary of the Invention

[0005] In order to overcome the shortcomings of the existing technology not mentioned above, the present invention aims to provide a caries detection device based on photometry and visual analysis. The detection device contains three platforms equivalent to three modules, namely, a structured light scanning module, a visual recognition module, and a light sectioning roughness detection module. The tooth surface contour and parameter model can be established through the collaborative processing of multi-source data. Combined with big data machine learning and artificial intelligence analysis, the multi-source data model obtained by scanning can be analyzed and diagnosed to determine whether caries occurs on the scanned tooth surface and locate the caries position.

[0006] In order to achieve the above object, the technical solution of the present invention is:

[0007] A dental caries detection device based on photometry and visual analysis, comprising a first macro camera 2 and a second macro camera 3, which are symmetrically arranged about a light source strip 1 and whose optical axes intersect, and are all fixedly connected to a first rotating platform 5;

[0008] Two fill lights 7 are distributed on both sides of the third macro camera 8 and all three are fixed on the second rotating platform 10;

[0009] The line laser generator 12 and the microscope camera 13 are fixed on both sides of the third rotating platform 15;

[0010] The first rotating platform 5, the second rotating platform 10, and the third rotating platform 15 are respectively hinged to the first rotating hinge support 4, the second rotating hinge support 9, and the third rotating hinge support 14 fixed on the detection base 17;

[0011] The first rotating platform 5, the second rotating platform 10, and the third rotating platform 15 are respectively connected to the first servo motor 6, the second servo motor 11, and the third servo motor 16 fixed on the detection base 17;

[0012] The detection base 17 is connected to the handle 21 through the sliding track 18 and the rod body 26, and the sliding position of the detection base 17 is controlled by the fourth servo motor 19; a protective cover 20 is provided on the top of the rod body 26;

[0013] The handle 21 and the rod body 26 are both equipped with built-in wires, and communicate with the signal processing system 22 via a signal line 25 for electrical signals; the display screen 23 communicates with the signal processing system 22 via electrical signals; and the signal processing system 22 is connected to a power supply 24 .

[0014] The focal lengths of the first macro camera 2 , the second macro camera 3 , and the third macro camera 8 are all smaller than or equal to half of the mouth opening, and automatic zooming can be achieved.

[0015] The fourth servo motor 19 is a linear servo motor.

[0016] The material of the protective cover 20 is a material with high light transmittance and high hardness.

[0017] The surface of the gripping area at the bottom of the handle 21 is made of flexible material.

[0018] The rod body 26 is made of clean and non-toxic materials, including composite resin and glass.

[0019] The signal processing system 22 has built-in tooth surface three-dimensional point cloud computing and deep learning software; the relative position of the macro camera and the target tooth surface is determined by the tooth surface three-dimensional point cloud computing to perform high-precision recognition and scanning of the tooth surface.

[0020] Compared with the prior art, the present invention has the following beneficial technical effects:

[0021] 1. The present invention is equipped with three rotating platforms, which respectively house a structured light scanning module, a visual recognition module, and a light section roughness detection module. By utilizing optical measurement and visual analysis, the risk of ionizing radiation is reduced, and the applicable population of the original dental caries detection method is expanded, especially for children and pregnant women, thus reducing the delay of the disease and enabling patients to receive effective and timely treatment.

[0022] The structured light scanning module includes a first macro camera and a second macro camera arranged in a directional arrangement of light strip sources and an intersecting binocular model. The light strip illuminates the jaw surface, the macro camera recognizes the light strip, and a signal processing system analyzes the light strip signal to obtain a high-precision three-dimensional structural model of each tooth surface. The visual recognition module includes a third macro camera that illuminates the jaw surface with a fill light. The third macro camera captures the image, and the signal processing system analyzes the pixels to obtain high-precision image color data of each tooth surface. Through pixel-level extraction and data analysis, subtle color changes on the tooth surface can be effectively and intuitively identified. The light section detection module uses a line laser to emit a vertical light strip to illuminate the target tooth surface. The reflected light passes through the lens and forms an image on the surface of the micro camera. An automatic sliding bracket, controlled by a servo motor, performs small, omnidirectional, and constant-speed movements to form a scanning structure. The micro camera transmits the captured continuous light strip signal to the signal processing center through stroboscopic light, which processes the signal to obtain the tooth surface roughness distribution.

[0023] 2. The test results are displayed on the display screen 23 in intuitive text, showing the location of the caries, the specific tooth surface, the depth of the caries, etc., without the need for professional analysis and judgment, and can be used by home residents for self-examination, facilitating the early detection and prevention of caries.

[0024] 3. The main part of the device has a simple structure, and the required equipment is small in size and low in cost. While improving the accuracy of the results and promoting the visualization of the results, it overcomes the problems of large size and complex operation of medical imaging equipment such as X-rays, promotes the application and promotion of the instrument, meets the daily oral health examination needs of ordinary residents, realizes home self-examination, and also improves the current situation of insufficient supply of imaging equipment in hospitals. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 It is a structural diagram of an embodiment of the present invention.

[0026] Figure 2 It is a front view schematic diagram of a working end of one embodiment of the present invention.

[0027] Figure 3 It is a right side schematic diagram of an embodiment of the present invention.

[0028] Figure 4 It is a left side schematic diagram of an embodiment of the present invention.

[0029] Figure 5 It is a schematic longitudinal section diagram of the structured light module of the present invention.

[0030] Figure 6 This is a flow chart of the present invention for realizing caries detection, location and severity analysis functions based on photometry and visual analysis. DETAILED DESCRIPTION

[0031] To facilitate understanding of the embodiments of the present invention, the present invention is further explained below with reference to the accompanying drawings and embodiments.

[0032] To better understand, Figure 1-5 As shown, the dental caries detection device based on photometry and visual analysis includes:

[0033] Reference Figure 1 、 Figure 2 、 Figure 3 A dental caries detection device based on photometry and visual analysis includes a first macro camera 2 and a second macro camera 3, which are symmetrically arranged about a light source band 1 and whose optical axes intersect, and all three are fixedly connected to a first rotating platform 5; the light band source 1 can project a stripe grating onto the target tooth surface, and the light band source has a certain light power to meet the needs of use under conventional lighting conditions.

[0034] The focal lengths of the first macro camera 2, the second macro camera 3, and the third macro camera 8 are all less than or equal to half of the mouth opening, and can achieve automatic zooming to achieve high-resolution image capture of the tooth.

[0035] The left and right optical axes of the first macro camera 2 and the second macro camera 3 intersect, ensuring that they can capture the maximum range of the fringe grating within the detection area.

[0036] Two fill lights 7 are distributed on both sides of the third macro camera 8 and all three are fixed on the second rotating platform 10; the line laser generator 12 and the microscope camera 13 are fixed on both sides of the third rotating platform 15; the fill lights 7 can project soft white light on the target tooth surface to prevent the light color from interfering with the tooth surface color data processing.

[0037] The line laser 12 can project a single light strip onto the surface of the tooth being tested; the microscopic camera 13 continuously collects the reflected light strips through stroboscopic light and transmits them to the signal processing system 22 through the built-in wires.

[0038] The first rotating platform 5 , the second rotating platform 10 and the third rotating platform 15 are respectively hinged to the first rotating hinge support 4 , the second rotating hinge support 9 and the third rotating hinge support 14 fixed on the detection base 17 .

[0039] The first rotating platform 5, the second rotating platform 10, and the third rotating platform 15 are respectively connected to the first servo motor 6, the second servo motor 11, and the third servo motor 16 fixed on the detection base 17; the first rotating platform 5, the second rotating platform 10, and the third rotating platform 15 all include rotating shafts, and the first servo motor, the second servo motor, and the third servo motor control the angles through the rotating shafts to which they are respectively connected, thereby controlling the first rotating platform 5, the second rotating platform 10, and the third rotating platform 15 to adjust the angle of the measuring device.

[0040] Reference Figure 4 、 Figure 5 The detection base 17 is connected to the handle 21 through the sliding track 18 and the rod body 26, and the sliding position of the detection base 17 is controlled by the fourth servo motor 19; the top of the rod body 26 is connected to the protective cover 20, and the protective cover 20 can be removed.

[0041] The handle 21 and the rod body 26 are both equipped with built-in wires, and communicate with the signal processing system 22 via a signal line 25 for electrical signals; the display screen 23 communicates with the signal processing system 22 via electrical signals; and the signal processing system 22 is connected to a power supply 24 .

[0042] The fourth servo motor 19 is a linear servo motor, which can control the detection base 17 to move linearly in the sliding track 18 .

[0043] The material of the protective cover 20 is a material with high light transmittance and high hardness, which can ensure that it is not easy to be scratched during use.

[0044] The surface of the gripping area at the bottom of the handle 21 is made of flexible material, which can improve grip comfort.

[0045] The rod body 26 is made of clean and non-toxic materials, including composite resin and glass, and can be used in the oral cavity.

[0046] The signal processing system 22 includes built-in three-dimensional point cloud computing and deep learning software for the tooth surface. This software uses the three-dimensional point cloud computing to determine the relative position of the macro camera and the target tooth surface, enabling high-precision identification and scanning of the tooth surface. Specifically, after the macro camera automatically focuses and the target tooth surface is clear within its field of view, it simultaneously captures a set of images of the caries lesion on the target tooth surface. A function is then used to determine the distance between the caries lesion and the camera. This is then subjected to linear regression analysis to eliminate outliers. Finally, the average value of the valid data is calculated as the distance between the caries lesion and the camera. Deep learning uses big data to deeply learn caries images and understand their characteristics. Through multi-source data processing, the geometric curves, color distribution, and roughness distribution of the tooth surface are extracted and detected. This multi-source data is then screened and judged by artificial intelligence to accurately determine whether the target tooth surface has caries.

[0047] The display screen 23 has high precision, high resolution, and original color display, so as to facilitate the display of data and images.

[0048] Reference Figure 6 , the working principle of the present invention is:

[0049] The three platforms are equivalent to three modules, including a structured light scanning module, a visual recognition module, and a light section roughness detection module. The structured light scanning module illuminates the maxillary face through a directionally arranged light source strip 1. The first macro camera 2 and the second macro camera 3 arranged in an intersecting binocular model recognize the light strip, and the signal processing system 22 analyzes the light strip signal. The first rotating platform 5, on which the light source strip 1 and the first and second macro cameras 2 and 3 are located, can be rotated under the control of a first servo motor 6 and supported by a rotating first rotating hinge support 4 to adjust the angle of the structured light scanning measurement device, thereby achieving the maximum range of acquisition of high-precision three-dimensional structural models of each tooth surface.

[0050] The visual recognition module illuminates the jaw surface with white light through the left and right fill lights 7 to reduce the interference of the ambient light color on the tooth surface color; the third macro camera 8 captures the image, and the signal processing system 22 performs RGB three-channel component analysis on the pixels; the second rotating platform 10 where the above-mentioned fill lights 7 and the third macro camera 8 are located can be rotated under the control of the second servo motor 11 and supported by the second rotating hinge support 9 to adjust the angle of the visual recognition device to achieve the maximum range of acquisition of the color distribution of each tooth surface.

[0051] The light sectioning roughness detection module emits a single vertical light strip through the line laser emitter 12 to illuminate the target tooth surface. The reflected light is imaged on the surface of the microscope camera 13 after reflection from the tooth surface. The microscope camera 13 continuously collects the reflected light strip signal through stroboscopic light, and the signal processing system 22 processes the light strip signal. The third rotating platform 15 where the above-mentioned line laser emitter 12 and the microscope camera 13 are located can be rotated under the control of the third servo motor 16 and the support of the third rotating hinge support 14 to adjust the angle of the light sectioning roughness detection device to achieve the maximum range of acquisition of the target tooth surface roughness distribution.

[0052] like Figure 1 As shown, the first rotating platform, the second rotating platform and the third rotating platform where the above three modules are located can slide up and down by the sliding track 18 with the detection base 17 as a unit, and the signal is collected by slightly adjusting the measuring position.

[0053] like Figure 1-5 As shown, when in use, the working end is placed in the patient's mouth. The structured light scanning system, visual recognition system, and light section detection system at the working end each emit and receive light signals, which are then transmitted to the signal processing system 22 via built-in wires. The outer side of the working end is provided with a rod 26 and a handle 21. The gripping area of ​​the handle is made of latex material to improve grip comfort. The signal processing device 22 then processes and analyzes the signals to obtain the detection results, which are displayed on the display screen 23.

[0054] like Figure 6 As shown, big data machine learning learns caries samples, and through the collaborative processing of multi-source data, it can establish a tooth surface parameter model data surface and mark the caries sample training features, process the caries samples, build a machine learning diagnosis model, grasp the characteristics of caries and establish a caries model; on the above basis, the caries detection system outputs a detection signal, and the structured light module, visual recognition module and light section detection module respectively send out characteristic light signals, perform contour scanning, image acquisition and surface microscopy on the target tooth surface to achieve three-dimensional reconstruction, pixel analysis and roughness characterization of the target tooth surface, extract the geometric surface, color distribution and roughness distribution of the detected tooth surface, and through the collaborative processing of multi-source data, the target tooth surface contour and parameter data surface can be established, and then the data of each sampling point on the surface can be calculated by weight network and artificial intelligence analysis, and compared with the caries model established by big data machine learning, the measured tooth surface can be effectively diagnosed, and caries information such as the tooth position, detailed location, and caries depth of the caries can be obtained.

[0055] In summary, the present invention projects linear structured light onto the target tooth surface through a light band source, scans the detected tooth surface as a whole through a sliding track and a rotating platform, and obtains a three-dimensional contour surface of the detected tooth surface through contour scanning and three-dimensional reconstruction; projects white light onto the target tooth surface through a fill light, visually recognizes the tooth surface by a third macro camera, and shoots the entire tooth surface through a sliding track and a rotating platform; and through RGB three-channel component analysis of the recognized pixels, the subtle surface color changes of early caries can be characterized by the proportion of heterochromatic pixels obtained through visual recognition, and the distribution of the degree of heterochromaticity of the tooth surface can be obtained; and the tooth surface can be scanned by a light sectioning method. The surface emits linear light bands, and in conjunction with the sliding track and rotating platform, it can perform microscopic identification of the overall roughness of the tooth surface and obtain the roughness distribution of the tested tooth surface. After obtaining various parameters of the tooth surface, a unified coordinate conversion is performed, and the camera coordinates are converted into world coordinates. Then, after deep learning and data processing by the information control center, a multi-source data surface is established. The geometric characteristics of this surface are the results of structured light scanning. The surface curvature can characterize the size and degree of caries cavities, and each point on the surface contains tooth surface heterochromaticity and roughness characteristics. Through neural networks and machine learning, a diagnostic weight network containing three factors of surface curvature, heterochromaticity, and roughness is established, and a diagnostic threshold is established. By performing weighted network calculation and artificial intelligence analysis on the data of each sampling point on the surface and comparing it with the established threshold, the tested tooth surface can be effectively diagnosed. By outputting the problematic sampling point and its surrounding data, the tooth position, detailed location, and caries depth of the caries lesion can be obtained, and the basic situation of caries can be simply and intuitively displayed, effectively improving the accuracy and visualization of caries detection, especially shallow caries detection. Secondly, since the detection medium is light rather than radiation, the potential psychological and physical harm caused by ionizing radiation is avoided, and the scope of applicable population of the dental caries detection device is expanded.

Claims

1. A dental caries detection device based on photometry and visual analysis, characterized in that: include: A first macro camera (2) and a second macro camera (3) are symmetrically arranged about the light source strip (1) and have intersecting optical axes, and are all fixedly connected to the first rotating platform (5); Two fill lights (7) are distributed on both sides of the third macro camera (8), and all three are fixed on the second rotating platform (10); The line laser generator (12) and the microscopic camera (13) are fixed on both sides of the third rotating platform (15); The first rotating platform (5), the second rotating platform (10), and the third rotating platform (15) are respectively hinged to the first rotating hinge support (4), the second rotating hinge support (9), and the third rotating hinge support (14) fixed on the detection base (17); The first rotating platform (5), the second rotating platform (10), and the third rotating platform (15) are respectively connected to a first servo motor (6), a second servo motor (11), and a third servo motor (16) fixed on a detection base (17); The detection base (17) is connected to the handle (21) through a sliding track (18) and a rod body (26), and the sliding position of the detection base (17) is controlled by a fourth servo motor (19); a protective cover (20) is provided on the top of the rod body (26); The handle (21) and the rod body (26) are both built with wires, and communicate with the signal processing system (22) through the signal line (25) for electrical signals; the display screen (23) communicates with the signal processing system (22) through electrical signals; the signal processing system (22) is connected to the power supply (24); The focal lengths of the first macro camera (2), the second macro camera (3), and the third macro camera (8) are all less than or equal to half the mouth opening, and automatic zooming is possible; The fourth servo motor (19) is a linear servo motor; The first macro camera (2) and the second macro camera (3) illuminate the jaw surface through the light source band (1), the two macro cameras recognize the light band, and the signal processing system (22) analyzes the light band signal, thereby obtaining a high-precision three-dimensional structural model of each tooth surface; The third macro camera (8) illuminates the jaw surface through the fill light (7), and the third macro camera (8) captures images. The signal processing system (22) analyzes the pixels, thereby obtaining high-precision image color data of each tooth surface. Through pixel-level extraction and data analysis, minute color changes on the tooth surface can be effectively and intuitively identified; The line laser generator (12) emits a vertical light strip to illuminate the target tooth surface, and the reflected light passes through the lens and forms an image on the surface of the micro camera (13). The detection base (17) moves in a small amplitude and at a constant speed in all directions under the control of the servo motor to form a scanning structure. The micro camera (13) transmits the captured continuous light strip signal to the signal processing system (22) through stroboscopic scanning, and the signal processing system (22) processes the signal to obtain the tooth surface roughness distribution.

2. The dental caries detection device based on photometry and visual analysis according to claim 1, characterized in that: The material of the protective cover (20) is a material with high light transmittance and high hardness.

3. The dental caries detection device based on photometry and visual analysis according to claim 1, characterized in that: The surface of the gripping area at the bottom of the handle (21) is made of flexible material.

4. The dental caries detection device based on photometry and visual analysis according to claim 1, characterized in that: The rod body (26) structural part is made of clean and non-toxic materials, including composite resin and glass.

5. The dental caries detection device based on photometry and visual analysis according to claim 1, characterized in that: The signal processing system (22) has built-in tooth surface three-dimensional point cloud computing and deep learning software; The relative position of the macro camera and the target tooth surface is determined through three-dimensional point cloud computing on the tooth surface, so that the tooth surface can be identified and scanned with high precision.

Images