CN113288014A

CN113288014A - Capsule endoscope system

Info

Publication number: CN113288014A
Application number: CN202110756854.2A
Authority: CN
Inventors: 崔大祥; 周诚; 史楠清; 沈琦; 彭家伟; 梁辉
Original assignee: Shanghai Jiaotong University; Shanghai National Engineering Research Center for Nanotechnology Co Ltd
Current assignee: Shanghai Jiaotong University; Shanghai National Engineering Research Center for Nanotechnology Co Ltd
Priority date: 2021-07-05
Filing date: 2021-07-05
Publication date: 2021-08-24

Abstract

The invention discloses a capsule endoscope system, which comprises a capsule endoscope and a targeting probe which is matched with the capsule endoscope for use and can specifically identify gastrointestinal diseases, wherein the capsule endoscope comprises a white light illumination module, an ultraviolet light illumination module, a near infrared light illumination module, an image acquisition module, a wireless image transmission module, an image processing and identification module based on deep learning, a capsule endoscope in-vivo position control module and the like; the target probe can specifically identify and adsorb to a specific lesion, and can emit fluorescence under the irradiation of a specific light source of the capsule endoscope, so that the specific lesion part is obviously different from a normal part, the target imaging is realized, the existing detection instrument based on the surface morphology is upgraded to the specific detection instrument based on the fluorescence, the defect that the existing endoscope easily leaks to detect diseases which do not cause obvious morphology change on the surface of the gastrointestinal tract is effectively overcome, and a new system for specific detection and specific treatment is provided for the gastrointestinal tract diseases.

Description

Capsule endoscope system

Technical Field

The invention relates to the field of medical instruments, in particular to a capsule endoscope system.

Background

Early screening of gastrointestinal diseases has very important significance, the treatment survival rate of early patients is far higher than that of middle and late patients, and the effective early screening can greatly improve the survival rate of related diseases. However, the early discovery rate of gastrointestinal diseases in China is very low, and most of gastrointestinal serious diseases are diagnosed at the late stage of the disease, which directly results in higher mortality rate. For example, the survival rate of early stage gastric cancer patients after 5 years of treatment is 84-99%, for advanced stage gastric cancer patients, the survival rate after 5 years of treatment is only 30-40%, and for advanced stage gastric cancer patients, the survival rate after 5 years of treatment is lower than 24%, but the discovery rate of early stage gastric cancer in China is lower than 15%. The low early detection rate of gastrointestinal diseases is partly due to the fact that symptoms are not obvious in early stages, and a more important reason is the lack of generally accepted intestinal noninvasive examination instruments with high accuracy.

At present, instruments clinically used for gastrointestinal tract disease examination are mainly insertion endoscopes based on shape judgment, namely, the shapes of gastrointestinal tract tissues are observed by naked eyes, suspected focuses are screened out, so that diagnosis is realized, meanwhile, the suspected focuses can be subjected to in vitro biopsy, and the method can be used for relevant treatment, such as laser ablation of polyps under an endoscope. However, the insertion endoscope can only use the obvious morphological change of the surface of the gastrointestinal tract as a basis, so that the micro focus such as lesion tissues below the mucosa and early tumors is difficult to find, and missed diagnosis or misdiagnosis is easy to cause. Besides the conventional insertion type endoscope, the insertion type examination instrument with the catheter can also be used for examining gastrointestinal tract diseases, such as pigment endoscope, ultrasonic endoscope, amplification endoscope, narrow-band imaging endoscope, confocal endoscope, cell endoscope and the like. Compared with the traditional insertion type endoscope, the more advanced endoscopes can realize the identification of gastrointestinal tract tiny lesions. However, the above inserted instruments are not suitable for people suffering from heart diseases, vascular diseases, poor lung function, low liver function, gastric perforation, erosive esophagitis, peritonitis, ascites, severe abdominal distension, cervical vertebra trauma, spinal deformity and the like, and can cause serious complications of some patients. In addition, the invasive detection causes great psychological and physiological discomfort to the examinee in the detection process, which is also the reason that most people choose to perform the examination only when the gastrointestinal tract has obvious discomfort, and do not perform the examination in daily physical examination.

In addition to the invasive detection instruments, the examination of the gastrointestinal tract may be performed by a series of non-invasive methods such as optical, acoustic, and magnetic fields, for example, gastrointestinal tract radiography detection, gastric ultrasound, nuclear magnetic resonance, and CT detection. The gastrointestinal tract radiography detection technology, the nuclear magnetic resonance and CT detection method can detect large gastrointestinal tract lesions, can not accurately detect small-size lesions, and has certain limitation. Since the stomach is a cavity, many air bubbles exist inside the stomach, which will affect the result of the ultrasound examination. Obviously, although the methods are widely applied, the methods are not good choices for gastrointestinal tracts and cannot become conventional gastrointestinal tract detection methods.

The emerging capsule type endoscope can also be used for examination of gastrointestinal diseases, and instructions and image data are transmitted in a wireless communication mode, so that the examination of the gastrointestinal tract can be completed without dragging a catheter, and the patient can be painless. After the examinee takes the capsule endoscope, the examination starts, and the examination ends as the capsule endoscope runs out of power. Because the capsule endoscope is disposable, cross infection is not caused, and the capsule endoscope is a great development of a gastrointestinal tract disease noninvasive detection technology. However, due to the limited power of the small battery, the capsule endoscope is exhausted before being expelled and cannot examine the entire digestive tract. Meanwhile, the capsule endoscope cannot get rid of the defect of 'only morphology theory' of the insertion endoscope, cannot diagnose the defect of pathological changes which do not cause obvious morphology change on the surface layer of the gastrointestinal tract, is difficult to find the pathological changes at folds and collapse parts of the gastrointestinal tract, cannot be matched with biopsy, and easily causes missed diagnosis or misdiagnosis. Meanwhile, the existing commercial capsule endoscope does not have a treatment function, and most of the capsule endoscopes with the treatment function are focused on releasing in-vivo drugs through the capsule endoscope, namely, after the capsule endoscope enters the body, the capsule endoscope is stimulated to release the drugs stored in the capsule endoscope, and the main problem at present cannot be solved.

Therefore, there is no ideal instrument for early screening of gastrointestinal lesions. In view of the current situation, there is a need to provide a non-invasive gastrointestinal disease screening and diagnosing apparatus which is suitable for the public and widely accepted, and can perform specific targeted accurate detection and diagnosis on gastrointestinal lesions without causing wound and pain to patients no matter whether the gastrointestinal lesions bring the morphological changes on the surface of the gastrointestinal tract, and meanwhile, has a certain therapeutic function, so as to solve the technical problems.

Disclosure of Invention

The invention aims to provide a capsule endoscope system which can carry out high-accuracy noninvasive examination on gastrointestinal diseases.

Yet another object of the present invention is to: the invention relates to a using method of a capsule endoscope system.

The capsule endoscope system comprises a capsule endoscope part and a targeting probe which is matched with the capsule endoscope part for use and can specifically identify gastrointestinal diseases, wherein the capsule endoscope part at least comprises a white light illumination module, an ultraviolet light illumination module, a near infrared light illumination module, an image acquisition module, a wireless image transmission module, an image processing and identifying module based on deep learning, a power supply module, a capsule endoscope in-vivo position control module, a signal processing module and a wireless communication module;

the targeting probe can specifically identify gastrointestinal tract diseases and at least comprises one of various substances which can be targeted and adsorbed to specific lesions, such as antibodies corresponding to proteins corresponding to genes highly expressed in cancerous tissues;

the capsule endoscope part is powered by the power module, the image acquired by the image acquisition module is processed by the signal processing module, and the image signal is transmitted to the image processing and identification module based on deep learning through the wireless communication module for identification and analysis.

On the basis of the scheme, the targeting probe can be coupled with a molecular contrast agent, when the molecular contrast agent is coupled with the targeting probe and used as a detector, the molecular contrast agent and the targeting probe are adsorbed to a specific pathological change part together, and then the pathological change part can generate fluorescence under the excitation of a specific light source of the capsule endoscope, so that the fluorescent probe is different from a normal part and other types of pathological changes, the targeting imaging is realized, and an image is acquired through an image acquisition module of the capsule endoscope and transmitted to the outside of a body. The method is convenient for doctors to judge whether the specific lesion exists through the images, so that the existing detection instrument based on the surface morphology is upgraded to a specific detection instrument based on fluorescence, the detection accuracy of gastrointestinal tract lesions is greatly improved, and a novel method for specific detection of gastrointestinal tract diseases is provided. Alternatively, the first and second electrodes may be,

on the basis of the scheme, the target probe can be coupled with the light absorption and heat production substance, when the light absorption and heat production substance is coupled with the target probe and used as a therapeutic apparatus, the light absorption and heat production substance absorbs light and then converts the light absorption and heat production substance into heat energy under the irradiation of the specific light source of the capsule endoscope, so that the lesion part generates local high temperature, local lesion tissues can be killed, and targeted photothermal ablation for specific lesions is realized. Provides a feasible method for the specific treatment of gastrointestinal tract disease patients by adopting the system of the invention. Alternatively, the first and second electrodes may be,

based on the above scheme, the substance capable of targeting and adsorbing to a specific lesion in the targeting probe is an antibody to a protein corresponding to a gene highly expressed in a cancerous tissue.

Furthermore, specific targeting probes corresponding to different lesions can be prepared, so that the probes can specifically target and identify the corresponding specific lesions. Specific targeting probes corresponding to different lesions can be prepared, so that the probes can specifically target and identify the corresponding specific lesions, such as helicobacter pylori, precancerous lesions, early tumors and the like.

On the basis of the scheme, the ultraviolet light illuminating module is arranged on the capsule endoscope and used for illuminating specific lesions, such as chronic atrophic gastritis, intestinal glandular metaplasia, atypical hyperplasia and the like, an autofluorescence image of the local lesion can be obtained through the image acquiring module of the capsule endoscope, and related images are transmitted to the outside of a body. Can be convenient for the doctor to judge relevant pathological changes, and can improve the accuracy of capsule endoscope detection of such specific pathological changes compared with visible light illumination.

On the basis of the scheme, the near-external light illumination module is arranged on the capsule endoscope and is used for matching with a targeting probe to perform specific targeting diagnosis and treatment, for example, a molecular contrast agent is excited to emit fluorescence, and a light source is provided for a light absorption and heat production substance to emit heat. The near-external light illumination module can greatly improve the accuracy of detecting specific lesions.

On the basis of the scheme, the image processing and identifying module based on deep learning can process and identify conventional gastrointestinal tract images, can process and identify gastrointestinal tract fluorescent images, and is also provided with a plurality of training models to match with different specific targeting probes to perform specific identification on the fluorescent images of specific lesions, so that the current situation that the conventional optical images can only be identified by the conventional image processing and identifying algorithm based on the capsule endoscope is improved, and the detection accuracy of the related specific lesions is improved.

Compared with the prior art, the invention has the beneficial effects that: the capsule endoscope system has the functions of a conventional capsule endoscope system, and can enable a focus area to have higher identification degree and accurately and specifically judge which lesion is by combining the capsule endoscope and the targeting probe under the cooperation of a specific light source, so that the existing detection instrument based on surface morphology is upgraded into a specific detection instrument based on fluorescence, a novel method for specifically detecting gastrointestinal diseases is provided, and the discovery rate of the gastrointestinal diseases is effectively improved. Meanwhile, local photothermal ablation treatment can be carried out on a specific lesion part, the problem that the existing capsule endoscope cannot be used for treatment is solved, a new instrument for specific treatment is provided for gastrointestinal tract diseases, and the application range of the capsule endoscope is effectively enlarged.

Drawings

FIG. 1 is a schematic view of a probe assembly of the capsule endoscopic system of the present invention;

FIG. 2 is a schematic view of a capsule endoscope composition of the capsule endoscope system of the present invention;

FIG. 3 is a schematic diagram of the image processing and recognition module based on deep learning of the capsule endoscopic system of the present invention;

FIG. 4 is a flow chart of a conventional detection of the capsule endoscopic system of the present invention as a detector;

FIG. 5 is a flow chart of the specific detection of the capsule endoscopic system of the present invention as a detector;

FIG. 6 is a flow chart of the operation of the capsule endoscopic system of the present invention as a therapy specific for a therapeutic device;

the reference numbers in the figures illustrate:

1-a targeting probe;

2-capsule endoscope;

201-capsule shaped package shell;

202-white light lighting module; 203-ultraviolet lighting module; 204-near infrared illumination module; 205-image acquisition module; 206-Power supply Module; 207-Wireless image transfer Module;

208-signal processing module; 209-capsule endoscope position control module;

a wireless communication module 210;

and 3, an image processing and identifying module based on deep learning.

Detailed Description

In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and should not be taken to be limiting.

Referring to fig. 1 to 3, the present invention provides a capsule endoscope system, which includes a targeting probe 1 capable of specifically recognizing gastrointestinal diseases, an image processing and recognition module 3 based on deep learning, a capsule-shaped packaging housing 201, and a white light illumination module 202, an ultraviolet light illumination module 203, a near-infrared light illumination module 204, an image acquisition module 205, a power supply module 206, a wireless image transmission module 207, a signal processing module 208, a capsule endoscope position control module 209 and a wireless communication module 210 disposed therein.

The targeting probe 1 can be selected and prepared for different lesions, so that the probe can be specifically targeted and adsorbed to a specific lesion. For example: selecting antibodies to proteins corresponding to the genes AF054162, AF054163 and/or AF054164 highly expressed in the precancerous lesion tissue for specific recognition of precancerous lesions such as chronic atrophic gastritis, intestinal metaplasia, atypical hyperplasia; selecting urease antibody and CagA antibody as the antibody targeting helicobacter pylori; selecting a BRCAA1 antibody, a HAI-178 humanized antibody, a PLCe1 antibody, and/or a gastrin G-17 antibody for discovery of early stage gastric cancer; the CD44-V6 antibody was selected to discover gastric cancer stem cells.

The probe and the molecular contrast agent can be near-infrared fluorescent probes and are coupled, so that the near-infrared fluorescent probes are absorbed to specific pathological change parts along with the probe in a targeted mode, at the moment, under the irradiation of a near-infrared light source of the capsule endoscope, the near-infrared fluorescent probes can emit fluorescence, the specific pathological change parts are obviously different from normal parts and other types of pathological changes, after related pictures are wirelessly transmitted to the outside of a body through the capsule endoscope, whether the specific pathological changes exist or not can be judged through analyzing the pictures.

The probe and the light-absorbing and heat-generating substance can be gold nano-star which are coupled, so that the gold nano-star is adsorbed to a specific pathological change part together with the probe in a targeted manner, and the gold nano-star can absorb light and generate heat under the irradiation of a near infrared light source of the capsule endoscope, so that the specific pathological change part generates local high temperature to be killed, and targeted photo-thermal ablation treatment is realized.

The white light illumination module 202 emits visible light in the same direction as the lens, has a visible wavelength range of 380nm-780nm, preferably 380nm, and is used for illuminating conventional lesions with obvious appearance characteristics and easy identification in gastrointestinal tracts, such as precancerous lesion gastric ulcer, gastric polyp and the like, and the lesions can be identified by acquiring conventional images.

The ultraviolet light illumination module 203 emits ultraviolet light in the same direction as the lens, the wavelength range of the ultraviolet light is 10nm to 380nm, preferably 230nm, and the ultraviolet light illumination module is used for illuminating some lesions which can generate autofluorescence images under the irradiation of the ultraviolet light, such as precancerous atrophic gastritis, intestinal metaplasia, atypical hyperplasia and the like, and the lesions can be identified and judged through the autofluorescence images.

The near infrared light illuminating module 204 emits near infrared light in the same direction as the lens, the wavelength range of the near infrared light is 780nm-2526nm, preferably 780nm, the near infrared light in this waveband avoids the main absorption regions of most endogenous substances such as oxyhemoglobin, deoxyhemoglobin, water and the like, so the penetration depth in the biological tissue is larger and can reach 10cm, and the fluorescence in this waveband is less affected by the background of the biological tissue. The near-infrared light illumination module needs to be matched with a targeting probe for use, and is mainly used for exciting the near-infrared fluorescent probe to emit light and exciting the gold nano-star to emit heat.

The image processing and recognition module 3 based on deep learning can be based on various peripherals, such as a desktop computer, a notebook computer, a tablet computer, a mobile phone and the like, the desktop computer is selected in the embodiment, the desktop computer has storage and display functions, and is provided with various image processing models based on deep learning, so that the conventional gastrointestinal tract pictures can be automatically recognized, and besides, the fluorescence image of a specific lesion can be specifically processed and recognized by matching with different specific targeting probes in a targeted manner, so that the gastrointestinal tract lesion condition can be automatically analyzed.

The capsule-shaped encapsulating shell 201 is made of a biocompatible material and is used for tightly encapsulating all parts of the capsule endoscope.

The image acquisition module 205, including an image sensor and a micro lens, is mainly used to acquire images of the gastrointestinal tract under different light sources, including conventional images and fluorescence images, and provide directions for the light sources.

The wireless image transmission module 207 is mainly used for wirelessly transmitting the image acquired by the image acquisition module to the outside of the body for reference of a doctor and for active analysis and identification of the image processing and identification module based on deep learning.

The power module 206 is mainly used for supplying power to the capsule endoscope, and may be a battery or a wireless coil based on the electromagnetic resonance inductive coupling principle.

The capsule endoscope internal position control module 209 may be an autonomous mechanical movement of the capsule endoscope, or may be a passive movement of the capsule endoscope operated by the outside, for example, the capsule endoscope is controlled by a magnetic control method, preferably by a magnetic control method, that is, the capsule endoscope includes a small permanent magnet, a support table on which the examinee lies outside, and an external magnetic control system for controlling the small permanent magnet in the capsule endoscope, which may be a large permanent magnet or an electromagnetic field. The position control module is mainly used for assisting the movement of the capsule endoscope and controlling the position of the capsule endoscope in the gastrointestinal tract so as to carry out fixed-point imaging and fixed-point photothermal ablation treatment.

The signal processing module 208, which is mainly an MCU in the capsule endoscope, is used to control the working states of the other modules of the capsule endoscope according to the preset requirements and the additionally added command requirements.

The wireless communication module 210 is mainly used for communicating with the outside so as to control the turning on and off of each light source and the image acquisition module from the outside.

Three procedures of diagnosis and treatment of gastrointestinal diseases on a subject by the capsule endoscope system according to the present embodiment will be described in detail below.

The flow chart of the routine inspection is shown in fig. 4. The capsule endoscope is taken by the examinee, and the white light module is started under the auxiliary control of the position control module along with the peristalsis of the gastrointestinal tract to collect the conventional images of the gastrointestinal tract. The doctor and the image processing and identifying module based on deep learning respectively identify the images transmitted by the wireless transmission, when the doctor finds that the images have suspected lesions, or the image processing and identifying module automatically identifies the suspected lesions and sends a prompt to the doctor through a preset program, the doctor can stop the capsule at a specific part through the position control module and acquire more images, or convert white light into ultraviolet light to acquire more information of the images according to the requirement, and comprehensively and perfectly analyze the suspected lesions. After the detection is finished, the white light illumination module is turned on again, the capsule endoscope continues moving along with the peristalsis of the gastrointestinal tract to continue detecting the gastrointestinal tract, and after the capsule is discharged out of the body of the examinee, the detection is finished.

As shown in fig. 5, when a specific lesion needs to be examined, the examinee first takes a targeting probe for the specific lesion, the probe is coupled with a near-infrared fluorescent probe, and then takes a capsule endoscope after a certain time interval, the capsule endoscope moves along with the gastrointestinal tract, and under the auxiliary control of a position control module, the images of the gastrointestinal tract are acquired, and at this time, the white light illumination module and the ultraviolet light illumination module do not need to be turned on, but only the near-infrared light illumination module needs to be turned on. The physician can determine whether the lesion exists according to the existence of fluorescence in the returned image, and record the position of the lesion. The image processing and recognition module based on deep learning trained by the detection mode can automatically analyze and detect the returned image, when a fluorescence signal is captured, a prompt is sent out through a preset program to determine that the lesion exists in the gastrointestinal tract of the examined person, and the position of the capsule endoscope at the moment, namely the position of the lesion, is recorded. The capsule endoscope then continues to move with the peristaltic movement of the gastrointestinal tract, and the detection of the gastrointestinal tract continues. When the images returned in the whole process have no fluorescence signals, the gastrointestinal tract is basically free from the pathological changes, and the detection is finished when the capsule is discharged out of the body by the examinee.

As shown in fig. 6, when diagnosis, detection and treatment are required for a specific lesion, a subject first takes a target probe for the specific lesion, the probe is coupled with a near-infrared fluorescent probe and a gold nanostar, and takes a capsule endoscope after a certain time interval, the capsule endoscope collects images of the gastrointestinal tract along with the peristalsis of the gastrointestinal tract under the auxiliary control of a position control module, and only the near-infrared illumination module is turned on at this time. The doctor can judge whether the pathological changes exist according to the fact that whether the returned images have fluorescence or not, and when the fluorescence images are found, the position control module is operated, so that the near infrared light of the capsule endoscope always faces the pathological changes, the gold nanostars continuously absorb light and generate heat, local pathological change tissues marked by the probe are killed, and targeted photo-thermal ablation treatment for specific pathological changes is achieved. Meanwhile, the image processing and recognition module based on deep learning trained by the detection mode can automatically analyze and detect the returned image, and when a fluorescence signal is captured, a prompt can be sent to a doctor through a preset program to assist the doctor in checking. And then the capsule endoscope continues moving along with the peristalsis of the gastrointestinal tract, the gastrointestinal tract is continuously detected, and the detection is finished after the capsule is discharged out of the body of the examinee.

The above description is only an example of the present invention and should not be construed as limiting the present invention, and any modification, equivalent replacement, and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A capsule endoscope system comprises a capsule endoscope part and is characterized by also comprising a targeting probe which is matched with the capsule endoscope part and can specifically identify gastrointestinal tract diseases;

the capsule endoscope part comprises a white light illumination module, an ultraviolet light illumination module, a near infrared light illumination module, an image acquisition module, a wireless image transmission module, a power supply module, an image processing and identifying module based on deep learning, a capsule endoscope in-vivo position control module, a signal processing module and a wireless communication module;

the targeting probe capable of specifically recognizing the gastrointestinal tract diseases at least comprises one of various substances capable of being targeted and adsorbed to specific lesions;

the image acquired by the image acquisition module is processed by the signal processing module, and the image signal is transmitted to the image processing and recognition module based on deep learning through the wireless communication module for recognition and analysis.

2. The capsule endoscopic system of claim 1, wherein the targeting probe is coupled to the molecular contrast agent, and when the molecular contrast agent is coupled to the targeting probe and adsorbed to the specific lesion site, the molecular contrast agent can cause the lesion site to generate fluorescence under the excitation of the specific light source of the capsule endoscope, so as to distinguish the lesion site from the normal site and other types of lesions, thereby realizing targeted imaging, and acquiring and transmitting the image to the outside of the body through the image acquisition module of the capsule endoscope.

3. The capsule endoscope system of claim 1, wherein the targeting probe is coupled with a light-absorbing and heat-generating substance, and after the light-absorbing and heat-generating substance is coupled with the targeting probe and absorbed to a specific lesion together, the light-absorbing and heat-generating substance absorbs light and then converts the light into heat energy under the irradiation of a specific light source of the capsule endoscope, so that the lesion generates local high temperature, and targeted photothermal ablation for the specific lesion is realized.

4. The capsule endoscope system of any one of claims 1 to 3, wherein the targeting probe comprises an antibody capable of targeting a protein corresponding to a gene highly expressed in a cancerous tissue as a substance adsorbed to a specific lesion, and the antibody comprises: antibodies to proteins corresponding to genes AF054162, AF054163 and/or AF054164 highly expressed in pre-cancerous diseased tissue; urease antibody and CagA antibody are used as antibodies targeting helicobacter pylori; BRCAA1 antibody, HAI-178 humanized antibody, PLCe1 antibody, and/or gastrin G-17 antibody, to discover early stage gastric cancer; CD44-V6 antibody to discover gastric cancer stem cells.

5. The capsule endoscopic system of claim 4, wherein the specific targeting probe is prepared for different lesions, so that the specific targeting probe can specifically target and identify the corresponding specific lesions.

6. The capsule endoscope system according to claim 1, wherein the ultraviolet illumination module is mounted on the capsule endoscope, illuminates a specific lesion, and obtains an autofluorescence image of a local portion of the lesion by an image acquisition module of the capsule endoscope, and transmits the relevant image to the outside of the body.

7. The capsule endoscopic system of claim 1, wherein the near-external light illumination module is mounted on the capsule endoscope and used for specific targeting diagnosis and treatment in cooperation with the targeting probe, exciting the molecular contrast agent to emit fluorescence, and providing a light source for the light-absorbing and heat-generating substance to emit heat.

8. The capsule endoscopic system of claim 1, wherein said capsule endoscopic portion is divided into an in vivo-accessible capsule endoscope and an out-of-body CPU-controlled deep learning based image processing and recognition module.

9. The capsule endoscopic system of claim 1 or 8, wherein the body-accessible capsule endoscope is enclosed by a biocompatible capsule-like enclosure consisting of a white light illumination module, an ultraviolet light illumination module, a near infrared light illumination module, an image acquisition module, a wireless image transmission module, a power supply module, a signal processing module, a capsule endoscope position control module, and a wireless communication module.

10. The capsule endoscope system of claim 1 or 8, wherein the image processing and recognition module based on deep learning is arranged outside the body, is controlled by a CPU, stores, calls and analyzes images, processes and recognizes gastrointestinal tract fluorescent images in addition to conventional gastrointestinal tract images, and has a plurality of training models to match different specific targeting probes to perform specific recognition on fluorescent images of specific lesions.