Target detection method for fast region search based on deep learning

Abstract

The invention discloses a target detection method for fast area search based on deep learning, which relates to the technical field of target detection and comprises the following implementation steps: configuring IP address information of a target detection remote control area server; entering a multi-feature detection end, collecting multi-feature information of the target object in real time, performing serial number arrangement, combining the serial number with the multi-feature information, and comparing in real time, so as to identify that the target object of the corresponding serial number is abnormal; entering a condition early warning end, and judging whether target detection abnormity occurs in a target detection process or not in real time in combination with an illumination condition, a background environment and a shooting angle; and entering a detection marking end, carrying out deviation calculation according to the position and size of the defect standard, and carrying out correction and compensation in time. According to the deep learning-based fast region search target detection method, the situation of missing detection or false detection during target detection is avoided, and whether deviation occurs in the position and size identification process of the target object is detected in real time.

Description

Technical Field

[0001] The present invention relates to the technical field of target detection, and in particular to a target detection method based on deep learning and rapid area search. Background Art

[0002] Fast regional search target detection based on deep learning is a method that uses deep learning technology to achieve efficient and accurate target detection. Target detection is the process of automatically identifying and locating specific objects in images or videos. The fast regional search target detection method based on deep learning aims to quickly find the location and category of the target object from the image through a deep learning model.

[0003] At present, there are some shortcomings in target detection based on deep learning fast regional search: 1. Due to the diversity of targets during target detection, some are large and some are small. For small-sized target objects in the image, it is easy to miss or misdetect. Because the feature information contained in small targets is limited, it is difficult to be accurately identified and located, which not only reduces the accuracy of regional search, but also reduces the accuracy and efficiency of target detection; 2. Since the environmental parameters of target detection are constantly changing, especially under different lighting conditions, background environments or shooting angles, the target object may not be accurately detected, resulting in poor target detection accuracy; 3. In the defect detection of the target object, the detected target position must be accurately marked. When the position and size information of each defect is quickly marked, the recognition of the position and size may be biased, and it is impossible to correct and compensate in time, resulting in limitations in target detection.

[0004] Therefore, a new target detection method based on deep learning and fast area search is proposed to solve the above problems. Summary of the invention

[0005] The main purpose of the present invention is to provide a target detection method for fast area search based on deep learning to solve the problems raised in the above background.

[0006] To achieve the above object, the technical solution adopted by the present invention is: a target detection method for fast area search based on deep learning, the method comprising the following implementation steps: Step 1: Configure the target detection remote control area server IP address information; Step 2: Enter the multi-feature detection terminal, collect multi-feature information of the target in real time, arrange the sequence number, compare the sequence number and multi-feature information in real time, identify the abnormality of the target with the corresponding sequence number, and if it is abnormal, re-detect the target and issue a voice alarm to remind the abnormality; Step 3: Enter the conditional warning terminal, collect the environmental parameters in the target detection process in real time, set the standard environmental parameters, and combine the lighting conditions, background environment and shooting angle to determine in real time whether there is any target detection abnormality in the target detection process; Step 4: Enter the detection and marking end, receive the target detection results in real time, mark the defects of the target object according to the serial number, and calculate the deviation according to the position and size of the defect marking. If a deviation occurs, correct it in time.

[0007] The multi-feature detection terminal includes a feature acquisition module, a sequence number arrangement module and an information comparison module; The feature acquisition module includes an information acquisition unit and a multi-feature recording unit; The information acquisition unit is used to set the regional search range and collect multi-feature information of the target object in real time through the data acquisition instrument, and the multi-feature information includes large targets, small targets, target positions and target shapes within the regional search range; The multi-feature recording unit is used to record the multi-feature information of the acquisition target object in real time through a data recorder.

[0008] The sequence number arrangement module includes a standard information unit and a sequence number arrangement unit; The standard information unit is used to set standard information for target detection according to the regional search range, and the standard information includes large targets, small targets, target positions and target shapes captured within the target search range; The serial number arrangement unit is used to arrange the multiple feature information in ascending order using Arabic numerals.

[0009] The information comparison module includes an information comparison unit, an abnormality warning unit and a duplication detection unit; The information comparison unit is used to set the multi-feature information of the collected target object as A=(a1, a2, ..., an), set the standard information of target detection as B=(b1, b2, ..., bn), and calculate the information difference value through a calculation formula, which is as follows: ; in, and Point Sets Hedianji The corresponding point in If the characteristic data value If it is equal to 0, it means that the information comparison is normal. If it is not equal to 0, it means that the information comparison is abnormal; The abnormal warning unit is used to report the abnormal information to the system and send out a voice alarm. The repeated detection unit is used to receive the information comparison abnormality signal, and collect the multi-feature information of the target object for a second time to re-compare the information until the information comparison is normal.

[0010] The condition warning terminal includes an environment acquisition module, a standard environment module and a comparison and judgment module; The environment acquisition module includes an illumination parameter unit, an environment parameter unit and a shooting angle unit; The illumination parameter unit is used to collect the illumination intensity of the corresponding area search range in real time through the light sensor; The environmental parameter unit is used to collect the temperature value of the corresponding area search range in real time through the temperature sensor; The shooting angle unit is used to capture the target detection shooting angle within the corresponding area search range in real time through a camera.

[0011] The standard environment module is used to set standard illumination intensity parameters, standard temperature parameters and standard shooting angle parameters.

[0012] The comparison and judgment module includes a one-by-one comparison unit, a comprehensive judgment unit and an impact warning unit; The one-to-one comparison unit is used to calculate the difference between the illumination intensity parameter and the illumination intensity, the standard temperature parameter and the temperature value, and the standard shooting angle parameter and the target detection shooting angle one by one, and record them in real time through a data recorder; The comprehensive judgment unit is used to combine the difference results of the illumination intensity parameter and the illumination intensity, the standard temperature parameter and the temperature value, and the standard shooting angle parameter and the target detection shooting angle to judge in real time whether the target detection abnormality occurs during the target detection process. The judgment method is as follows: The difference results are arranged in sequence by Arabic numerals and recorded as D1, D2 and D3 in sequence. If D1, D2 and D3 are all less than or equal to 0.5 and greater than or equal to -0.5, it is judged that the target detection process is normal; if not, it is judged that the target detection process is abnormal; The impact warning unit is used to report the system to issue a voice alarm reminder when the target detection is abnormal.

[0013] The detection and marking end includes a data receiving module, a defect marking module, a deviation calculation module and a correction and compensation module; The data receiving module includes a normal recording unit and a defect recording unit; The normal recording unit is used to receive the target detection judgment result in real time through the data receiver, and define the target object whose target detection result is normal as a normal target; The defect recording unit is used to define a target object whose detection result is abnormal as an abnormal target.

[0014] The defect annotation module includes a defect serial number unit and a defect annotation unit; The defect serial number unit is used to automatically identify the serial number corresponding to the abnormal target; The defect marking unit is used to automatically mark the defect position according to the sequence number of the abnormal target.

[0015] The deviation calculation module is used to calculate the comprehensive deviation value of the defect position in real time and automatically mark it from multiple angles. , the calculation method is as follows: The actual position point of the defect position is automatically marked as (x measured, y measured, z measured), and the standard marking position coordinates are (x theoretical, y theoretical, z theoretical), then: Deviation along the x-axis: Δx = x measured − x theoretical; Deviation along the y-axis: Δy = y measured − y theoretical; Deviation along the z-axis: Δz = z measured − z theoretical; Combined with multi-angle calculation of defect location for automatic marking of comprehensive deviation value : ; The correction compensation module includes a correction compensation adjustment unit and a correction self-tracking unit; The correction compensation adjustment unit is used to automatically compensate and adjust the defect position marking parameters according to the deviation value; The corrected self-tracking unit is used to calculate the difference between the marked parameters after automatic compensation adjustment and the marked parameters before automatic compensation adjustment. If the difference is equal to 0, it means that the corrected compensation is abnormal. If the difference is not equal to 0, it means that the corrected compensation is normal.

[0016] The present invention has the following beneficial effects: 1. In the present invention, by setting a multi-feature detection terminal, when detecting a target based on a fast regional search of deep learning, by combining the serial number and the multi-feature information for real-time comparison, it is identified whether the target information of the corresponding serial number is abnormal, so that when detecting a target, it is possible to combine the multi-features of the target and the capture points arranged by the serial number, thereby avoiding missed detection or false detection during target detection, and even small targets can be accurately identified, which not only increases the accuracy of the regional search, but also improves the accuracy and efficiency of target detection; 2. In the present invention, by setting a conditional warning terminal, during the target detection based on deep learning rapid area search, by combining the lighting conditions, background environment and shooting angle around the target object, it is judged in real time whether the target detection abnormality occurs during the target detection process, so that the environmental parameters of the target detection are constantly changing, especially under different lighting conditions, background environments or shooting angles, the target object can be accurately detected, thereby improving the accuracy of target detection; 3. In the present invention, by setting a detection and marking end, during target detection based on deep learning rapid area search, the deviation is calculated in real time according to the position and size of the defect standard, so that in the process of target detection based on deep learning rapid area search, the position and size information of each defect can be accurately and quickly marked, and whether there is a deviation in the position and size identification process of the target object can be detected in real time, and when a deviation occurs, correction and compensation can be made in time to reduce the limitations existing in target detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 This is an overall flow chart of a target detection method for fast area search based on deep learning of the present invention; Figure 2 A schematic diagram of the architecture of a multi-feature detection end of a target detection method for fast area search based on deep learning according to the present invention; Figure 3 A schematic diagram of the structure of a conditional warning terminal of a target detection method for rapid area search based on deep learning according to the present invention; Figure 4 This is a schematic diagram of the architecture of the detection and annotation end of a target detection method for fast area search based on deep learning in the present invention. DETAILED DESCRIPTION

[0018] In order to make the technical means, creative features, objectives and effects achieved by the present invention easy to understand, the present invention is further explained below in conjunction with specific implementation methods.

[0019] Embodiment 1 Please refer to Figure 1 to Figure 2 As shown: A target detection method for fast area search based on deep learning, the method includes the following implementation steps: Step 1: Configure the target detection remote control area server IP address information; Step 2: Enter the multi-feature detection terminal, collect multi-feature information of the target in real time, arrange the sequence number, compare the sequence number and multi-feature information in real time, identify the abnormality of the target with the corresponding sequence number, and if it is abnormal, re-detect the target and issue a voice alarm to remind the abnormality; Step 3: Enter the conditional warning terminal, collect the environmental parameters in the target detection process in real time, set the standard environmental parameters, and combine the lighting conditions, background environment and shooting angle to determine in real time whether there is any target detection abnormality in the target detection process; Step 4: Enter the detection and marking end, receive the target detection results in real time, mark the defects of the target object according to the serial number, and calculate the deviation according to the position and size of the defect marking. If a deviation occurs, correct it in time.

[0020] The multi-feature detection terminal includes a feature acquisition module, a sequence number arrangement module and an information comparison module; The feature acquisition module includes an information acquisition unit and a multi-feature recording unit; The information collection unit is used to set the regional search range and collect multi-feature information of the target object in real time through the data collection instrument. The multi-feature information includes large targets, small targets, target positions and target shapes within the regional search range; The multi-feature recording unit is used to record the multi-feature information of the collected target object in real time through a data recorder, arrange the multi-feature information in sequence, combine the sequence number and the multi-feature information for real-time comparison, and identify whether the target object information with the corresponding sequence number is abnormal.

[0021] The sequence number arrangement module includes a standard information unit and a sequence number arrangement unit; The standard information unit is used to set standard information for target detection according to the regional search range, and the standard information includes large targets, small targets, target positions and target shapes captured within the target search range; The serial number arrangement unit is used to arrange the multiple feature information in ascending order using Arabic numerals.

[0022] The information comparison module includes an information comparison unit, an abnormal warning unit and a duplicate detection unit; The information comparison unit is used to set the multi-feature information of the collected target object as A=(a1, a2, ..., an), set the standard information of target detection as B=(b1, b2, ..., bn), and calculate the information difference value through the calculation formula. The calculation formula is as follows: ; in, and Point Sets Hedianji The corresponding point in If the characteristic data value If it is equal to 0, it means that the information comparison is normal. If it is not equal to 0, it means that the information comparison is abnormal; The abnormal warning unit is used to report abnormal information to the system and issue a voice alarm. The repeated detection unit is used to receive information comparison abnormal signals, and collect multi-feature information of the target object for a second time to re-compare the information until the information comparison is normal. By combining the multi-features of the target object and the capture points arranged in sequence, accurate detection of diverse targets can be achieved to avoid missed detection or false detection during target detection. Even small targets can be accurately identified, which not only increases the accuracy of regional search, but also improves the accuracy and efficiency of target detection.

[0023] Embodiment 2 Please refer to Figure 3 As shown: Based on the first embodiment, the condition warning terminal includes an environment acquisition module, a standard environment module and a comparison and judgment module; The environment acquisition module includes an illumination parameter unit, an environment parameter unit and a shooting angle unit; The illumination parameter unit is used to collect the illumination intensity of the corresponding area search range in real time through the light sensor; The environmental parameter unit is used to collect the temperature value of the corresponding area search range in real time through the temperature sensor; The shooting angle unit is used to capture the target detection shooting angle within the corresponding area search range in real time through the camera.

[0024] The standard environment module is used to set standard light intensity parameters, standard temperature parameters and standard shooting angle parameters. By real-time acquisition of the environmental parameters around the target during target detection and setting the standard environmental parameters for target detection, combined with the lighting conditions, background environment and shooting angle around the target, it is possible to determine in real time whether there are any target detection anomalies during target detection.

[0025] The comparison and judgment module includes a one-by-one comparison unit, a comprehensive judgment unit and an impact warning unit; The one-to-one comparison unit is used to calculate the difference between the illumination intensity parameter and the illumination intensity, the standard temperature parameter and the temperature value, and the standard shooting angle parameter and the target detection shooting angle one by one, and record them in real time through the data recorder; The comprehensive judgment unit is used to combine the difference results of the light intensity parameter and the light intensity, the standard temperature parameter and the temperature value, and the standard shooting angle parameter and the target detection shooting angle to judge in real time whether the target detection abnormality occurs during the target detection process. The judgment method is as follows: The difference results are arranged in sequence by Arabic numerals and recorded as D1, D2 and D3 in sequence. If D1, D2 and D3 are all less than or equal to 0.5 and greater than or equal to -0.5, it is judged that the target detection process is normal; if not, it is judged that the target detection process is abnormal; The impact warning unit is used to report abnormalities in target detection. The system issues a voice alarm reminder. By judging whether there are abnormalities in target detection during the target detection process, the environmental parameters of target detection are constantly changing. The accuracy of target detection can also be guaranteed, especially under different lighting conditions, background environments or shooting angles. The target object can be accurately detected, thereby improving the accuracy of target detection.

[0026] Embodiment 3 Please refer to Figure 4 As shown: Based on the first embodiment, the detection and marking end includes a data receiving module, a defect marking module, a deviation calculation module and a correction and compensation module; The data receiving module includes a normal recording unit and a defect recording unit; The normal recording unit is used to receive the target detection judgment result in real time through the data receiver, and define the target object whose target detection result is normal as a normal target; The defect recording unit is used to define a target object whose detection result is abnormal as an abnormal target.

[0027] The defect annotation module includes a defect serial number unit and a defect annotation unit; The defect serial number unit is used to automatically identify the serial number corresponding to the abnormal target; The defect marking unit is used to automatically mark the defect position according to the serial number of the abnormal target, mark the defect of the target object according to the serial number, calculate the deviation in real time according to the position and size of the defect standard, and make corrections and compensation in time if deviation occurs.

[0028] The deviation calculation module is used to calculate the comprehensive deviation value of the defect position in real time and automatically mark it from multiple angles. , the calculation method is as follows: The actual position point of the defect position is automatically marked as (x measured, y measured, z measured), and the standard marking position coordinates are (x theoretical, y theoretical, z theoretical), then: Deviation along the x-axis: Δx = x measured − x theoretical; Deviation along the y-axis: Δy = y measured − y theoretical; Deviation along the z-axis: Δz = z measured − z theoretical; Combined with multi-angle calculation of defect location for automatic marking of comprehensive deviation value : ; The correction compensation module includes a correction compensation adjustment unit and a correction self-tracking unit; The correction compensation adjustment unit is used to automatically compensate and adjust the defect position marking parameters according to the deviation value, and the size information is automatically adjusted according to the size of the target defect; The correction self-tracking unit is used to calculate the difference between the marking parameters after automatic compensation adjustment and the marking parameters before automatic compensation adjustment. If the difference is equal to 0, it means that the correction compensation is abnormal. If the difference is not equal to 0, it means that the correction compensation is normal. It can accurately mark the detected target objects and quickly mark the position and size information of each defect. It can also detect in real time whether there is any deviation in the position and size identification process of the target object, and make timely corrections and compensation when deviations occur.

[0029] In the present invention, a target detection method for fast regional search based on deep learning is provided. First, the IP address information of the target detection remote control regional server is configured; the multi-feature detection terminal is entered, and the multi-feature information of the target is collected in real time and arranged in sequence, and the sequence number and the multi-feature information are combined for real-time comparison to identify the abnormality of the target with the corresponding sequence number. If abnormal, the target detection is performed again, and a voice alarm is issued to remind the abnormality. The multi-feature information of the target is collected in real time, and the multi-feature information is arranged in sequence, and the sequence number and the multi-feature information are combined for real-time comparison to identify whether the target information with the corresponding sequence number is abnormal, so that when detecting the target, the multi-features of the target and the capture points arranged in sequence can be combined to achieve accurate detection of diverse targets, avoid missed detection or false detection during target detection, and even small targets can be accurately identified, which not only increases the accuracy of regional search, but also improves the accuracy and efficiency of target detection; the condition warning terminal is entered, and the environmental parameters in the target detection process are collected in real time, standard environmental parameters are set, and the lighting conditions, background environment and shooting angle are combined to judge in real time whether the target detection abnormality occurs in the target detection process, and the target detection abnormality is detected by collecting the surroundings of the target in the target detection process in real time. Environmental parameters, and set standard environmental parameters for target detection, combined with the lighting conditions, background environment and shooting angle around the target, to determine in real time whether target detection anomalies occur during the target detection process, so that the environmental parameters for target detection are constantly changing, and the accuracy of target detection can be guaranteed, especially under different lighting conditions, background environments or shooting angles, the target object can be accurately detected to improve the accuracy of target detection; enter the detection and marking end, receive the target detection results in real time, and mark the defects of the target according to the serial number, and calculate the deviation according to the position and size of the defect standard. If a deviation occurs, correct and compensate in time, by receiving the target detection results in real time, and marking the defects of the target according to the serial number, and calculating the deviation according to the position and size of the defect standard in real time. If a deviation occurs, correct and compensate in time, so that in the target detection process of fast regional search based on deep learning, not only can the detected target be accurately marked, and the position and size information of each defect can be quickly marked, but also whether there is a deviation in the position and size recognition process of the target can be detected in real time, and when a deviation occurs, correct and compensate in time, reducing the limitations of target detection.

[0030] The above shows and describes the basic principles and main features of the present invention and the advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments. The above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A target detection method for fast area search based on deep learning, characterized in that: The method comprises the following implementation steps: Step 1: Configure the target detection remote control area server IP address information; Step 2: Enter the multi-feature detection terminal, collect multi-feature information of the target in real time, arrange the sequence number, compare the sequence number and multi-feature information in real time, identify the abnormality of the target with the corresponding sequence number, and if it is abnormal, re-detect the target and issue a voice alarm to remind the abnormality; Step 3: Enter the conditional warning terminal, collect the environmental parameters in the target detection process in real time, set the standard environmental parameters, and combine the lighting conditions, background environment and shooting angle to determine in real time whether there is any target detection abnormality in the target detection process; Step 4: Enter the detection and marking end, receive the target detection results in real time, mark the defects of the target object according to the serial number, and calculate the deviation according to the position and size of the defect marking. If a deviation occurs, correct it in time.

2. The method according to claim 1, characterized in that: The multi-feature detection terminal includes a feature acquisition module, a sequence number arrangement module and an information comparison module; The feature acquisition module includes an information acquisition unit and a multi-feature recording unit; The information acquisition unit is used to set the regional search range and collect multi-feature information of the target object in real time through the data acquisition instrument, and the multi-feature information includes large targets, small targets, target positions and target shapes within the regional search range; The multi-feature recording unit is used to record the multi-feature information of the acquisition target object in real time through a data recorder.

3. The method according to claim 2, characterized in that: The sequence number arrangement module includes a standard information unit and a sequence number arrangement unit; The standard information unit is used to set standard information for target detection according to the regional search range, and the standard information includes large targets, small targets, target positions and target shapes captured within the target search range; The serial number arrangement unit is used to arrange the multiple feature information in ascending order using Arabic numerals.

4. The method according to claim 3, characterized in that: The information comparison module includes an information comparison unit, an abnormality warning unit and a duplication detection unit; The information comparison unit is used to set the multi-feature information of the collected target object as A=(a1, a2, ..., an), set the standard information of target detection as B=(b1, b2, ..., bn), and calculate the information difference value through a calculation formula, which is as follows: ； in, and Point Sets Hedianji The corresponding point in If the characteristic data value If it is equal to 0, it means that the information comparison is normal. If it is not equal to 0, it means that the information comparison is abnormal; The abnormal warning unit is used to report the abnormal information to the system and send out a voice alarm. The repeated detection unit is used to receive the information comparison abnormality signal, and collect the multi-feature information of the target object for a second time to re-compare the information until the information comparison is normal.

5. The method according to claim 1, characterized in that: The condition warning terminal includes an environment acquisition module, a standard environment module and a comparison and judgment module; The environment acquisition module includes an illumination parameter unit, an environment parameter unit and a shooting angle unit; The illumination parameter unit is used to collect the illumination intensity of the corresponding area search range in real time through the light sensor; The environmental parameter unit is used to collect the temperature value of the corresponding area search range in real time through the temperature sensor; The shooting angle unit is used to capture the target detection shooting angle within the corresponding area search range in real time through a camera.

6. The method according to claim 5, characterized in that: The standard environment module is used to set standard illumination intensity parameters, standard temperature parameters and standard shooting angle parameters.

7. The method according to claim 6, characterized in that: The comparison and judgment module includes a one-by-one comparison unit, a comprehensive judgment unit and an impact warning unit; The one-to-one comparison unit is used to calculate the difference between the illumination intensity parameter and the illumination intensity, the standard temperature parameter and the temperature value, and the standard shooting angle parameter and the target detection shooting angle one by one, and record them in real time through a data recorder; The comprehensive judgment unit is used to combine the difference results of the illumination intensity parameter and the illumination intensity, the standard temperature parameter and the temperature value, and the standard shooting angle parameter and the target detection shooting angle to judge in real time whether the target detection abnormality occurs during the target detection process. The judgment method is as follows: The difference results are arranged in sequence by Arabic numerals and recorded as D1, D2 and D3 in sequence. If D1, D2 and D3 are all less than or equal to 0.5 and greater than or equal to -0.5, it is judged that the target detection process is normal; if not, it is judged that the target detection process is abnormal; The impact warning unit is used to report the system to issue a voice alarm reminder when the target detection is abnormal.

8. The method according to claim 1, characterized in that: The detection and marking end includes a data receiving module, a defect marking module, a deviation calculation module and a correction and compensation module; The data receiving module includes a normal recording unit and a defect recording unit; The normal recording unit is used to receive the target detection judgment result in real time through the data receiver, and define the target object whose target detection result is normal as a normal target; The defect recording unit is used to define a target object whose detection result is abnormal as an abnormal target.

9. The method according to claim 8, characterized in that: The defect annotation module includes a defect serial number unit and a defect annotation unit; The defect serial number unit is used to automatically identify the serial number corresponding to the abnormal target; The defect marking unit is used to automatically mark the defect position according to the sequence number of the abnormal target.

10. The method according to claim 9, characterized in that: The deviation calculation module is used to calculate the comprehensive deviation value of the defect position in real time and automatically mark it from multiple angles. , the calculation method is as follows: The actual position point of the defect position is automatically marked as (x measured, y measured, z measured), and the standard marking position coordinates are (x theoretical, y theoretical, z theoretical), then: Deviation along the x-axis: Δx = x measured − x theoretical; Deviation along the y-axis: Δy = y measured − y theoretical; Deviation along the z-axis: Δz = z measured − z theoretical; Combined with multi-angle calculation of defect location for automatic marking of comprehensive deviation value : ； The correction compensation module includes a correction compensation adjustment unit and a correction self-tracking unit; The correction compensation adjustment unit is used to automatically compensate and adjust the defect position marking parameters according to the deviation value; The corrected self-tracking unit is used to calculate the difference between the marked parameters after automatic compensation adjustment and the marked parameters before automatic compensation adjustment. If the difference is equal to 0, it means that the corrected compensation is abnormal. If the difference is not equal to 0, it means that the corrected compensation is normal.

Images