WO2020082396A1 - 一种检测方法和检测装置 - Google Patents

一种检测方法和检测装置 Download PDF

Info

Publication number
WO2020082396A1
WO2020082396A1 PCT/CN2018/112272 CN2018112272W WO2020082396A1 WO 2020082396 A1 WO2020082396 A1 WO 2020082396A1 CN 2018112272 W CN2018112272 W CN 2018112272W WO 2020082396 A1 WO2020082396 A1 WO 2020082396A1
Authority
WO
WIPO (PCT)
Prior art keywords
processor
radiation
preset exposure
preset
test object
Prior art date
Application number
PCT/CN2018/112272
Other languages
English (en)
French (fr)
Inventor
王星泽
倪一帆
舒远
Original Assignee
合刃科技（深圳）有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 合刃科技（深圳）有限公司 filed Critical 合刃科技（深圳）有限公司
Priority to CN201880068619.2A priority Critical patent/CN111356914B/zh
Priority to PCT/CN2018/112272 priority patent/WO2020082396A1/zh
Publication of WO2020082396A1 publication Critical patent/WO2020082396A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
    • G01N21/84—Systems specially adapted for particular applications
    • G01N21/88—Investigating the presence of flaws or contamination

Definitions

  • the embodiments of the present application relate to the field of electronic technology, and in particular, to a detection method and a detection device.
  • the traditional detection method is to manually use a magnifying glass to detect electronic products. Because electronic products have the characteristics of compact structure and concentrated component settings, when using traditional detection methods to detect electronic products, the test results rely on the visual observation of the tester, efficiency Low accuracy is not guaranteed.
  • the embodiments of the present application provide a detection method and a detection device, which can improve detection efficiency and detection accuracy.
  • an embodiment of the present application provides a detection method, which is applied to an electronic device, the electronic device includes: a radiation emitting unit, a detector, a processor, and a memory, and a test object is placed on the radiation emitting unit and Between the detectors; the method includes the following steps:
  • the processor obtains N preset exposure durations corresponding to the measured object from the memory, where N is an integer equal to 2 or greater than 2, and the radiographic transmission pattern corresponding to any of the preset exposure durations includes One or more clear target sub-regions;
  • the processor obtains a radiation transmission diagram of the object to be measured corresponding to each preset exposure duration according to the radiation acquired by the detector in each of the N preset exposure durations;
  • the processor identifies each target sub-region in the radiographic transmission diagram, and obtains a fusion image of the test object according to the identified target sub-region;
  • the processor recognizes the fusion image according to a defect model, and obtains a detection result of the object to be tested.
  • the embodiment of the present application uses radiation to irradiate the test object to obtain multiple radiographic perspectives corresponding to different exposure durations, to identify the target sub-region in each radiation transmission diagram, to obtain a fusion image according to the identified target sub-region, and then to compare The fusion image is used for recognition, and the detection result of the measured object is obtained. Since the fused image is obtained from multiple clear target sub-regions, identifying the fused image according to the defect model is beneficial to improve the detection efficiency and detection accuracy.
  • the electronic device further includes an output unit
  • the method further includes:
  • the processor sends a first message to the output unit, where the first message carries the detection result
  • the output unit outputs the detection result according to the received first message.
  • the detection result includes: one or more of normal, damaged, burrs, scratches, crooked terminals, and missing pins;
  • the output unit may include a display screen, which may display the probability of various possible detection results, or the number of various detection results, or a combination of pictures representing the test object.
  • the location where the defect is detected is marked in the figure, and the detected defect is marked. This is helpful for visually displaying and observing the possible defects of the measured object.
  • the electronic device further includes a sampling unit, and before the processor acquires N preset exposure durations corresponding to the measured object from the memory, the method further includes :
  • the sampling unit recognizes the measured object, and sends the recognition result to the processor;
  • the processor obtains N preset exposure durations corresponding to the measured object from the memory according to the obtained recognition result.
  • the recognition result may be the type identifier of the test object
  • different types of test objects may correspond to different exposure durations
  • the category identification and the corresponding exposure duration information may be pre-stored in the database
  • the processor obtains the type of the test object After the identification, search the data stored in the database to obtain the N exposure durations corresponding to the test object.
  • the processor can directly obtain N preset exposure durations corresponding to the measured object.
  • the electronic device further includes a transmission unit; before the sampling unit identifies the object to be measured, the method further includes:
  • the transmission unit moves different objects to be detected to a detection position every preset time period, and the detection position is located between the radiation emission source and the detector.
  • Using this embodiment can automatically move a plurality of measured objects to the detection position one by one for detection, which is beneficial to improve the detection efficiency.
  • the method before the processor recognizes the fused image according to a defect model, the method further includes:
  • the processor performs defect data training and learning according to a preset defect training sample corresponding to the test object to obtain a defect model corresponding to the test object.
  • the defect model can be obtained by the processor through defect data training and learning according to the preset defect samples, or can be obtained from other devices.
  • the method before the processor obtains N preset exposure durations corresponding to the measured object from the memory, the method further includes:
  • the processor determines N preset exposure durations corresponding to the test object according to the structure and / or experiment of the test object, and saves the N preset exposure durations to the memory.
  • an embodiment of the present application provides a detection device, the detection device includes: a radiation emitting unit, a detector, a processor, and a memory, and a test object is placed between the radiation emitting unit and the detector ;
  • the radiation emitting unit is used to emit radiation, and the radiation is used to irradiate the object to be measured;
  • the detector is used to receive the radiation transmitted by the measured object
  • the processor is configured to obtain N preset exposure durations corresponding to the measured object from the memory, where N is an integer equal to or greater than 2, and any radiation corresponding to the preset exposure duration
  • the transmission image includes one or more clear target sub-regions
  • the processor is further configured to obtain the measured object corresponding to each preset exposure duration according to the radiation acquired by the detector in each of the N preset exposure durations Radiograph of
  • the processor is also used to identify each target sub-region in the radiographic transmission diagram, obtain a fusion image of the object under test based on the identified target sub-region, and identify the fusion image according to the defect model To obtain the detection result of the test object.
  • the embodiment of the present application irradiates the test object with radiation to obtain a plurality of radiographic perspectives corresponding to different exposure durations, and then identifies the target sub-region in each radiograph, obtains the fusion image according to the identified target sub-region, and then according to the defect model Recognize the fusion image and get the detection result of the measured object. Since the fused image is obtained from multiple clear target sub-regions, identifying the fused image according to the defect model is beneficial to improve the detection efficiency and detection accuracy.
  • the electronic device further includes: an output unit;
  • the processor is further configured to send a first message to the output unit, where the first message carries the detection result;
  • the output unit outputs the detection result according to the received first message.
  • the detection result includes: one or more of normal, damaged, burrs, scratches, crooked terminals, and missing pins;
  • the output unit may include a display screen, which may display the probability of various possible detection results, or the number of various detection results, or a combination of pictures representing the test object.
  • the location where the defect is detected is marked in the figure, and the detected defect is marked. This is helpful for visually displaying and observing the possible defects of the measured object.
  • the detection device further includes:
  • the sampling unit is used to identify the object to be tested before the processor obtains N preset exposure durations corresponding to the object from the memory, and send the recognition result to the processor .
  • the processor in terms of acquiring N preset exposure durations corresponding to the object to be measured from the memory, is specifically used to acquire from the memory based on the identification result obtained from the sampling unit. Describe the N preset exposure durations corresponding to the measured object.
  • the detection device further includes:
  • the transmission unit is used to move the different test object to a detection position every preset time period before the sampling unit recognizes the test object, the detection position is located at the radiation emission source and the Between the detectors.
  • Using this embodiment can automatically move a plurality of measured objects to the detection position one by one for detection, which is beneficial to improve the detection efficiency.
  • the processor is further configured to, before identifying the defect of the test object based on the defect model and the fusion image, correspond to the test object according to a preset Of the defect training samples for defect data training and learning to obtain the defect model corresponding to the test object.
  • the processor is further configured to, according to the structure of the test object and / or before acquiring N preset exposure durations corresponding to the test object from the memory Or test to determine the N preset exposure durations corresponding to the test object, and keep the N preset exposure durations in the memory.
  • FIG. 1 is a schematic diagram of an application scenario of an electronic device including a detection device provided by an embodiment of the present application
  • FIG. 2A is a schematic flowchart of a detection method provided by an embodiment of the present application.
  • 2B is a schematic flowchart of another detection method provided by an embodiment of the present application.
  • 2C is a schematic flowchart of another detection method provided by an embodiment of the present application.
  • 3A is a schematic structural diagram of a detection device provided by an embodiment of the present application.
  • 3B is a schematic structural diagram of another detection device provided by an embodiment of the present application.
  • 3C is a schematic structural diagram of another detection device provided by an embodiment of the present application.
  • the form and structure of the connector are ever-changing. With different application objects, frequency, power, application environment, etc., there are various types of connectors, but no matter what kind of connector, the current must be guaranteed during its normal operation. Smooth and continuous and reliable circulation.
  • the pin in the connector is an important structure of the connector. It is mainly used for electrical conduction and signal transmission. During the production process of the connector, when the pin feet are assembled, there are uneven forces due to the movement of the cylinder. In some cases, the pin feet may be tilted, which affects the final quality of the product. Therefore, it is necessary to test the pin feet of the connector.
  • the traditional detection method is that the tester uses a magnifying glass to check whether the connector is qualified. The visual intuition, the accuracy is low, and the accuracy is not guaranteed.
  • the embodiment of the present application irradiates the test object with radiation, obtains a plurality of target sub-regions using the radiation transmission maps corresponding to different exposure durations, obtains a fusion image of the test object according to the target sub-region, and then recognizes the fusion image according to the defect model, Obtain the test result of the test object. Since the fused image is obtained from multiple clear target sub-regions, identifying the fused image according to the defect model is beneficial to improve the detection efficiency and detection accuracy.
  • FIG. 1 is a schematic diagram of an application scenario of an electronic device including a detection device according to an embodiment of the present application.
  • the electronic device may include: a radiation emitting unit 101, a detector 103, a memory 104, and a processor 105, and the measured object 102 is placed between the radiation emitting unit 101 and the detector 103;
  • the radiation emitting unit 101 emits radiation to irradiate the object under test 102
  • the detector 103 generates an image corresponding to the object under test 102 according to the radiation passing through the object under test 102
  • the processor 105 acquires the object to be tested from the memory 104 102 corresponding to N preset exposure durations, where N is an integer equal to 2 or greater than 2, and the radiograph corresponding to any preset exposure duration includes one or more clear target sub-regions
  • the processor 105 according to the detector 103 Radiation acquired at each preset exposure duration in N preset exposure durations to obtain a radiograph of the test object 102 corresponding
  • FIG. 2A is a schematic flowchart of a detection method provided by an embodiment of the present application.
  • the detection method is applied to an electronic device.
  • the electronic device includes: a radiation emitting unit, a detector, a processor, and a memory. Between the radiation emitting unit and the detector; the method includes steps 201-204, as follows:
  • the processor obtains N preset exposure durations corresponding to the object to be measured from the memory, where N is an integer equal to 2 or greater than 2, and any radiation transmission map corresponding to the preset exposure duration includes One or more clear target sub-regions.
  • the processor can obtain a clear image of the structure G3 according to the radiation received by the detector, and the preset exposure time corresponding to the structure G3 is 9 seconds. Therefore, the three preset exposure durations corresponding to the object to be measured can be preset to be 2 seconds, 5 seconds, and 9 seconds, respectively.
  • the material at different positions of the measured object is the same, because there are differences in shape, structure, etc., such as different thickness, it will also make the different positions of the measured object at different exposure durations.
  • the definition of the image is different, and the images obtained according to different exposure time can detect different positions. Understandably, the preset duration of the test object can be determined by experiment, and the preset duration is saved in the memory after the determination. When detecting a certain test object, the preset corresponding to the test object is taken out from the memory duration.
  • the processor obtains a radiation transmission diagram of the object to be measured corresponding to each preset exposure duration according to the radiation acquired by the detector in each of the N preset exposure durations. .
  • the processor can obtain the radiation transmission maps corresponding to the measured object under different exposure durations.
  • the measured object 102 includes three different areas: area A1, area A2, and area An, where area A1 is the clearest in the radiographic transmission diagram when the exposure duration is T1, and area A2 is during the exposure duration
  • area A1 is the clearest in the radiographic transmission diagram when the exposure duration is T1
  • area A2 is during the exposure duration
  • the radiation transmission diagram corresponding to T2 is the clearest
  • the area An is clearest when the exposure duration is Tn.
  • the processor generates radiation transmission diagrams p1, p2, and p3 of the test object according to the radiation received by the detector.
  • the processor identifies each target sub-region in the radiographic transmission diagram, and obtains a fusion image of the test object according to the identified target sub-region.
  • the processor 105 Taking the measured object 102 in FIG. 1 as an example, the processor 105 generates the radiographic transmission patterns p1, p2, and p of the measured object according to the radiation received by the detector when the exposure time of the measured object 102 is T1, T2, and Tn respectively. After p3, the processor identifies the clear areas in the perspective views, ie the target sub-areas of each picture, for example, the target sub-area A1 from figure p1, the target sub-area A2 from figure p2, and the target Area An. Then the processor processes the extracted target sub-regions to obtain the fusion image p including A1, A2, and A3.
  • the test object when the test object is a connector, the materials of different structural parts on the connector are different, so that the imaging quality of the different structural parts on the connector is affected by the exposure time factor to a different degree, which can be expressed in brightness, contrast, edge
  • the area with the above parameters can be turned into a favorable area, and the radiation can be X-rays. X-rays irradiate the test object at different exposure times to obtain different images of the favorable area. The intersection of the favorable areas is the smallest and the union is the largest. The images are fused to obtain a clear image of each part as a fused image, which is used to further detect whether the connector is defective.
  • the processor recognizes the fusion image according to a defect model, and obtains a detection result of the test object.
  • the processor performs defect data training and learning according to a preset defect training sample corresponding to the test object to obtain a defect model corresponding to the test object.
  • a preset defect training sample corresponding to the test object
  • the image corresponding to the area including the defect may be displayed as the detection result. Understandably, it can be further reminded of the existence of defects by voice or icon or dynamic reminder picture.
  • the processor can use the deep learning neural network to automatically identify the defects of the measured object.
  • a large number of training samples are used to identify different defects of the test object. For example, more than 100 images are collected as training samples according to different defect types. The samples are classified and trained to obtain the test object defect image after training. During the detection, the image fusion image is compared with the pre-obtained test object defect image to identify possible defects.
  • the recognition results may include: Normal, damaged, burrs, scratches, crooked terminals, missing pins, etc.
  • the embodiment of the present application uses radiation to irradiate the test object to obtain multiple radiographic perspectives corresponding to different exposure durations, to identify the target sub-region in each radiation transmission diagram, to obtain a fusion image according to the identified target sub-region, and then to compare The fusion image is used for recognition, and the detection result of the measured object is obtained. Since the fused image is obtained from multiple clear target sub-regions, identifying the fused image according to the defect model is beneficial to improve the detection efficiency and detection accuracy.
  • FIG. 2B is a schematic flowchart of a detection method provided by another embodiment of the present application.
  • the detection method is applied to an electronic device.
  • the electronic device includes a radiation emitting unit, a detector, a processor, and a memory. Placed between the radiation emitting unit and the detector; the method includes steps 211-215, as follows:
  • the sampling unit recognizes the measured object, and sends the recognition result to the processor.
  • the sampling unit starts to detect the test object.
  • the test object as a connector for example, when the connector is placed between the radiation emitting unit and the detector After that, the sampling unit recognizes the object to be measured, and recognizes the object to be a connector.
  • the processor obtains N preset exposure durations corresponding to the measured object from the memory according to the obtained recognition result.
  • the N is an integer equal to 2 or greater than 2, and the radiograph corresponding to any preset exposure duration includes one or more clear target sub-regions.
  • the preset exposure duration corresponding to at least one detection object is stored in the memory, for example, N preset exposure durations corresponding to the connector are stored in the memory in advance, for example, N may be 3, and the preset exposure duration It can be 2 seconds, 5 seconds, and 9 seconds.
  • the structure can obtain a clear image at the time of exposure. For example, if the structure G1 can obtain a clear image of the structure G1 according to the radiation received by the detector when the exposure time is 2 seconds, the structure G1 corresponds to the preset The exposure time is 2 seconds.
  • the structure G2 can obtain a clear image of the structure G2 according to the radiation received by the detector when the exposure time of the structure G2 is 5 seconds, the preset exposure time corresponding to the structure G2 is 5 seconds. If the exposure time of the structure G3 is 9 seconds, the processor can obtain a clear image of the structure G3 according to the radiation received by the detector, and the preset exposure time corresponding to the structure G3 is 9 seconds. Therefore, the three preset exposure durations corresponding to the object to be measured can be preset to be 2 seconds, 5 seconds, and 9 seconds, respectively.
  • the material at different positions of the measured object is the same, because there are differences in shape, structure, etc., such as different thickness, it will also make the different positions of the measured object at different exposure durations.
  • the definition of the image is different, and the images obtained according to different exposure time can detect different positions. Understandably, the preset duration of the test object can be determined by experiment, and the preset duration is saved in the memory after the determination. When detecting a certain test object, the preset corresponding to the test object is taken out from the memory duration.
  • the processor obtains a radiation transmission diagram of the object to be measured corresponding to each preset exposure duration according to the radiation acquired by the detector in each of the N preset exposure durations. .
  • the processor can obtain the radiation transmission maps corresponding to the measured object under different exposure durations.
  • the measured object 102 includes three different areas: area A1, area A2, and area An, where area A1 is the clearest in the radiographic transmission diagram when the exposure duration is T1, and area A2 is during the exposure duration
  • area A1 is the clearest in the radiographic transmission diagram when the exposure duration is T1
  • area A2 is during the exposure duration
  • the radiation transmission diagram corresponding to T2 is the clearest
  • the area An is clearest when the exposure duration is Tn.
  • the processor generates radiation transmission diagrams p1, p2, and p3 of the test object according to the radiation received by the detector.
  • the processor identifies each target sub-region in the radiographic transmission diagram, and obtains a fusion image of the test object according to the identified target sub-region.
  • the processor 105 Taking the measured object 102 in FIG. 1 as an example, the processor 105 generates the radiographic transmission patterns p1, p2, and p of the measured object according to the radiation received by the detector when the exposure time of the measured object 102 is T1, T2, and Tn, respectively. After p3, the processor identifies the clear areas in the perspective views as the target sub-areas of each picture, such as identifying the target sub-area A1 from figure p1, identifying the target sub-area A2 from figure p2, and identifying the target sub-area from figure p3 Area An. Then the processor processes the extracted target sub-regions to obtain the fusion image p including A1, A2, and A3.
  • the test object when the test object is a connector, the materials of different structural parts on the connector are different, so that the imaging quality of the different structural parts on the connector is affected by the exposure time factor to a different degree, which can be expressed in brightness, contrast, edge
  • the area with the above parameters can be turned into a favorable area, and the radiation can be X-rays. X-rays irradiate the test object at different exposure times to obtain different images of the favorable area. The images are fused to obtain a clear image of each part as a fused image, which is used to further detect whether the connector is defective. 215.
  • the processor recognizes the fusion image according to a defect model, and obtains a detection result of the test object.
  • the processor performs defect data training and learning according to a preset defect training sample corresponding to the test object to obtain a defect model corresponding to the test object.
  • a preset defect training sample corresponding to the test object
  • the image corresponding to the area including the defect may be displayed as the detection result. Understandably, it can be further reminded of the existence of defects by voice or icon or dynamic reminder picture.
  • the processor can use the deep learning neural network to automatically identify the defects of the measured object.
  • a large number of training samples are used to identify different defects of the test object. For example, more than 100 images are collected as training samples according to different defect types. The samples are classified and trained to obtain the test object defect image after training. During the detection, the image fusion image is compared with the pre-obtained test object defect image to identify possible defects.
  • the recognition results may include: Normal, damaged, burrs, scratches, crooked terminals, missing pins, etc.
  • FIG. 2C is a schematic flowchart of a detection method provided by another embodiment of the present application.
  • the detection method is applied to an electronic device.
  • the electronic device includes a radiation emitting unit, a detector, a processor, and a memory.
  • the object is placed between the radiation emitting unit and the detector; the method includes steps 221-226, as follows:
  • the transmission unit transmits different objects to the detection position every preset time period, and the detection position is located between the radiation emission source and the detector.
  • the interval of the test object being conveyed by the conveyor belt is determined by the detection speed of the test object. It can be set that the detection result of the current test object comes out before the next test object is transmitted to the next test object Between the radiation emitting unit and the detector. For example, if the detection period of a test object is 15 seconds, the conveyor belt transfers another test object to the detection position every 15 seconds. It can be understood that different test objects have different transmission speeds corresponding to the conveyor belt. For example, if the detection time of the first type of test object is 15 seconds and the detection time of the second type of test object is 30 seconds, then the When a type of test object is detected, the conveyor belt transfers the next first type test object to the test position every 15 seconds.
  • the conveyor belt transfers the next second type of test object to the measured position every 30 seconds. It can be understood that the conveyor belt transmission speed may also use the detection time corresponding to the object with the longest detection time as the interval time between two adjacent objects to be conveyed by the conveyor belt.
  • the sampling unit recognizes the measured object, and sends the recognition result to the processor.
  • the sampling unit starts to detect the test object.
  • the test object as a connector for example, when the connector is placed between the radiation emitting unit and the detector After that, the sampling unit recognizes the object to be measured, and recognizes the object to be a connector.
  • the processor obtains N preset exposure durations corresponding to the measured object from the memory according to the obtained recognition result.
  • the N is an integer equal to 2 or greater than 2.
  • the radiographic map corresponding to any of the preset exposure duration includes one or more clear target sub-regions.
  • the preset exposure duration corresponding to at least one detection object is stored in the memory, for example, N preset exposure durations corresponding to the connector are stored in the memory in advance, for example, N may be 3, and the preset exposure duration It can be 2 seconds, 5 seconds, and 9 seconds.
  • the structure can obtain a clear image at the time of exposure. For example, if the structure G1 can obtain a clear image of the structure G1 according to the radiation received by the detector when the exposure time is 2 seconds, the structure G1 corresponds to the preset The exposure time is 2 seconds.
  • the structure G2 can obtain a clear image of the structure G2 according to the radiation received by the detector when the exposure time of the structure G2 is 5 seconds, the preset exposure time corresponding to the structure G2 is 5 seconds. If the exposure time of the structure G3 is 9 seconds, the processor can obtain a clear image of the structure G3 according to the radiation received by the detector, and the preset exposure time corresponding to the structure G3 is 9 seconds. Therefore, the three preset exposure durations corresponding to the object to be measured can be preset to be 2 seconds, 5 seconds, and 9 seconds, respectively.
  • the material at different positions of the measured object is the same, because there are differences in shape, structure, etc., such as different thickness, it will also make the different positions of the measured object at different exposure durations.
  • the definition of the image is different, and the images obtained according to different exposure time can detect different positions. Understandably, the preset duration of the test object can be determined by experiment, and the preset duration is saved in the memory after the determination. When detecting a certain test object, the preset corresponding to the test object is taken out from the memory duration.
  • the processor obtains a radiation transmission diagram of the object to be measured corresponding to each preset exposure duration according to the radiation acquired by the detector in each of the N preset exposure durations. .
  • the processor can obtain the radiation transmission maps corresponding to the measured object under different exposure durations.
  • the measured object 102 includes three different areas: area A1, area A2, and area An, where area A1 is the clearest in the radiographic transmission diagram when the exposure duration is T1, and area A2 is during the exposure duration
  • area A1 is the clearest in the radiographic transmission diagram when the exposure duration is T1
  • area A2 is during the exposure duration
  • the radiation transmission diagram corresponding to T2 is the clearest
  • the area An is clearest when the exposure duration is Tn.
  • the processor generates radiation transmission diagrams p1, p2, and p3 of the test object according to the radiation received by the detector.
  • the processor identifies each target sub-region in the radiographic transmission diagram, and obtains a fusion image of the test object according to the identified target sub-region.
  • the processor 105 Taking the measured object 102 in FIG. 1 as an example, the processor 105 generates the radiographic transmission patterns p1, p2, and p of the measured object according to the radiation received by the detector when the exposure time of the measured object 102 is T1, T2, and Tn, respectively. After p3, the processor identifies the clear areas in the perspective views as the target sub-areas of each picture, such as identifying the target sub-area A1 from figure p1, identifying the target sub-area A2 from figure p2, and identifying the target sub-area from figure p3 Area An. Then the processor processes the extracted target sub-regions to obtain the fusion image p including A1, A2, and A3.
  • the test object is a connector
  • the materials of different structural parts on the connector are different, so that the imaging quality of the different structural parts on the connector is affected by the exposure time factor to a different degree, which can be expressed in brightness, contrast, edge
  • the area with the above parameters can be turned into a favorable area, and the radiation can be X-rays. X-rays irradiate the test object at different exposure times to obtain different images of the favorable area. The images are fused to obtain a clear image of each part as a fused image, which is used to further detect whether the connector is defective.
  • the processor recognizes the fusion image according to a defect model, and obtains a detection result of the test object.
  • the transfer unit can automatically move multiple objects to the detection position one by one for detection, which is beneficial to improve the detection efficiency.
  • the electronic device further includes an output unit
  • the method further includes:
  • the processor sends a first message to the output unit, where the first message carries the detection result
  • the output unit outputs the detection result according to the received first message.
  • the output unit may include a display screen, which may display the probability of various possible detection results, or the number of various detection results, or a combination of pictures representing the test object.
  • the location where the defect is detected is marked in the figure, and the detected defect is marked. This is helpful for visually displaying and observing the possible defects of the measured object.
  • the terminal includes a hardware structure and / or a software module corresponding to each function.
  • the present application can be implemented in the form of hardware or a combination of hardware and computer software. Whether a function is executed by hardware or computer software driven hardware depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.
  • the embodiments of the present application may divide the functional unit of the terminal according to the above method example, for example, each functional unit may be divided corresponding to each function, or two or more functions may be integrated into one processing unit.
  • the above integrated unit can be implemented in the form of hardware or software function unit. It should be noted that the division of units in the embodiments of the present application is schematic, and is only a division of logical functions, and there may be other division manners in actual implementation.
  • FIG. 3A is a schematic structural diagram of a detection device according to an embodiment of the present application.
  • the detection device 310 includes: a radiation emitting unit 311, a detector 312, a processor 313, and a memory 314, and the object 300 is placed between the radiation emitting unit 311 and the detector 312;
  • the radiation emitting unit 311 is used to emit radiation, and the radiation is used to irradiate the object to be measured.
  • the detector 312 is used to receive the radiation transmitted by the object.
  • the processor 313 is configured to obtain N preset exposure durations corresponding to the measured object from the memory, where N is an integer equal to 2 or greater than 2, and any radiation transmission corresponding to the preset exposure duration
  • the graph includes one or more clear target sub-regions.
  • the processor 313 is further configured to obtain, according to the radiation acquired by the detector in each of the N preset exposure durations, the radiation of the object to be measured corresponding to each preset exposure duration Radiograph.
  • the processor can obtain a clear image of the structure G3 according to the radiation received by the detector, and the preset exposure time corresponding to the structure G3 is 9 seconds. Therefore, the three preset exposure durations corresponding to the object to be measured can be preset to be 2 seconds, 5 seconds, and 9 seconds, respectively.
  • the material at different positions of the measured object is the same, because there are differences in shape, structure, etc., such as different thickness, it will also make the different positions of the measured object at different exposure durations.
  • the definition of the image is different, and the images obtained according to different exposure time can detect different positions. Understandably, the preset duration of the test object can be determined by experiment, and the preset duration is saved in the memory after the determination. When detecting a certain test object, the preset corresponding to the test object is taken out from the memory duration.
  • the processor can obtain the radiation transmission maps corresponding to the measured object under different exposure durations.
  • the measured object 102 includes three different areas: area A1, area A2, and area An, where area A1 is the clearest in the radiographic transmission diagram when the exposure duration is T1, and area A2 is the exposure duration
  • area A1 is the clearest in the radiographic transmission diagram when the exposure duration is T1
  • area A2 is the exposure duration
  • the radiation transmission diagram corresponding to T2 is the clearest
  • the area An is clearest when the exposure duration is Tn.
  • the processor when the exposure duration of the test object 102 is T1, T2, and Tn, the processor generates radiation transmission diagrams p1, p2, and p3 of the test object according to the radiation received by the detector.
  • the processor 313 is also used to identify each target sub-region in the radiographic transmission diagram, obtain a fusion image of the object under test based on the identified target sub-region, and identify the fusion image according to the defect model, The detection result of the test object is obtained.
  • the processor 105 Taking the measured object 102 in FIG. 1 as an example, the processor 105 generates the radiographic transmission patterns p1, p2, and p of the measured object according to the radiation received by the detector when the exposure time of the measured object 102 is T1, T2, and Tn, respectively. After p3, the processor identifies the clear areas in the perspective views as the target sub-areas of each picture, such as identifying the target sub-area A1 from figure p1, identifying the target sub-area A2 from figure p2, and identifying the target sub-area from figure p3 Area An. Then the processor processes the extracted target sub-regions to obtain the fusion image p including A1, A2, and A3.
  • the test object when the test object is a connector, the materials of different structural parts on the connector are different, so that the imaging quality of the different structural parts on the connector is affected by the exposure time factor to a different degree, which can be expressed in brightness, contrast, edge
  • the area with the above parameters can be turned into a favorable area, and the radiation can be X-rays. X-rays irradiate the test object at different exposure times to obtain different images of the favorable area. The intersection of the favorable areas is the smallest and the union is the largest. The images are fused to obtain a clear image of each part as a fused image, which is used to further detect whether the connector is defective.
  • the processor performs defect data training and learning according to a preset defect training sample corresponding to the test object to obtain a defect model corresponding to the test object.
  • a preset defect training sample corresponding to the test object
  • the image corresponding to the area including the defect may be displayed as the detection result. Understandably, it can be further reminded of the existence of defects by voice or icon or dynamic reminder picture.
  • the embodiment of the present application uses radiation to irradiate the test object to obtain multiple radiographic perspectives corresponding to different exposure durations, to identify the target sub-region in each radiation transmission diagram, to obtain a fusion image according to the identified target sub-region, and then to compare The fusion image is used for recognition, and the detection result of the measured object is obtained. Since the fused image is obtained from multiple clear target sub-regions, identifying the fused image according to the defect model is beneficial to improve the detection efficiency and detection accuracy.
  • FIG. 3B is a schematic structural diagram of another detection device according to an embodiment of the present application. As shown in FIG. 3B, this embodiment, compared with the embodiment shown in FIG. 3A, the detection device 310 further includes a sampling unit 315.
  • the sampling unit 315 is used to identify the object to be measured 300 before the processor 313 acquires N preset exposure durations corresponding to the object to be measured 300 from the memory 314, and send the recognition result to the processor 313.
  • the sampling unit starts to detect the test object.
  • the test object as a connector for example, when the connector is placed between the radiation emitting unit and the detector After that, the sampling unit recognizes the object to be measured, and recognizes the object to be a connector.
  • the preset exposure duration corresponding to at least one detection object is stored in the memory, for example, N preset exposure durations corresponding to the connector are stored in the memory in advance, for example, N may be 3, and the preset exposure duration It can be 2 seconds, 5 seconds, and 9 seconds.
  • the structure can obtain a clear image at the time of exposure. For example, if the structure G1 can obtain a clear image of the structure G1 according to the radiation received by the detector when the exposure time is 2 seconds, the structure G1 corresponds to the preset The exposure time is 2 seconds.
  • the structure G2 can obtain a clear image of the structure G2 according to the radiation received by the detector when the exposure time of the structure G2 is 5 seconds, the preset exposure time corresponding to the structure G2 is 5 seconds. If the exposure time of the structure G3 is 9 seconds, the processor can obtain a clear image of the structure G3 according to the radiation received by the detector. The preset exposure time corresponding to the structure G3 is 9 seconds. Therefore, the three preset exposure durations corresponding to the object to be measured can be preset to be 2 seconds, 5 seconds, and 9 seconds, respectively.
  • the material at different positions of the measured object is the same, because there are differences in shape, structure, etc., such as different thickness, it will also make the different positions of the measured object at different exposure durations.
  • the definition of the image is different, and the images obtained according to different exposure time can detect different positions. Understandably, the preset duration of the test object can be determined by experiment, and the preset duration is saved in the memory after the determination. When detecting a certain test object, the preset corresponding to the test object is taken out from the memory duration.
  • the processor can directly obtain N preset exposure durations corresponding to the measured object.
  • FIG. 3C is a schematic structural diagram of another detection device according to an embodiment of the present application. As shown in FIG. 3C, this embodiment, compared with the embodiment shown in FIG. 3B, the detection device 310 further includes a transmission unit 316.
  • the transmission unit 316 is used to transmit different objects 300 to the detection position every preset time interval before the sampling unit 315 recognizes the object 300, and the detection position is located between the radiation emitting unit 311 and the detector 312.
  • the interval of the test object being conveyed by the conveyor belt is determined by the detection speed of the test object. It can be set that the detection result of the current test object comes out before the next test object is transmitted to the next test object Between the radiation emitting unit and the detector. For example, if the detection period of a test object is 15 seconds, the conveyor belt transfers another test object to the detection position every 15 seconds. It can be understood that different test objects have different transmission speeds corresponding to the conveyor belt. For example, if the detection time of the first type of test object is 15 seconds and the detection time of the second type of test object is 30 seconds, then the When a type of test object is detected, the conveyor belt transfers the next first type test object to the test position every 15 seconds.
  • the conveyor belt transfers the next second type of test object to the measured position every 30 seconds. It can be understood that the conveyor belt transmission speed may also use the detection period corresponding to the test object with the longest detection time as the interval time between two adjacent test objects on the conveyor belt.
  • the transfer unit can automatically move multiple objects to the detection position one by one for detection, which is beneficial to improve the detection efficiency.
  • the detection device may further include: an output unit; the processor is further configured to send a first message to the output unit, where the first message carries the detection result; the The output unit outputs the detection result according to the received first message.
  • the detection results include: one or more of normal, damaged, burrs, scratches, crooked terminals, missing pins; the output unit may include a display screen, which may display the probability of various possible detection results, Either the number of various detection results, or a picture representing the object to be tested, the location of the detected defect is marked in the figure, and the detected defect is marked. This is helpful for visually displaying and observing the possible defects of the measured object.
  • An embodiment of the present application further provides an electronic device including a detection device.
  • the detection device may be the detection device described in any of the foregoing embodiments.
  • An embodiment of the present application further provides a computer storage medium, wherein the computer storage medium stores a computer program for electronic data exchange, and the computer program causes the computer to execute part of any of the detection methods described in the above method embodiments All steps.
  • An embodiment of the present application further provides a computer program product, the computer program product includes a non-transitory computer-readable storage medium storing a computer program, the computer program causes the computer to perform any of the detections described in the above method embodiments Some or all steps of the method.
  • the disclosed device may be implemented in other ways.
  • the device embodiments described above are only schematic.
  • the division of the unit is only a logical function division.
  • there may be another division manner for example, multiple units or components may be combined or may Integration into another system, or some features can be ignored, or not implemented.
  • the displayed or discussed mutual coupling or direct coupling or communication connection may be indirect coupling or communication connection through some interfaces, devices or units, and may be in electrical or other forms.
  • the program may be stored in a computer-readable memory, and the memory may include: a flash disk , Read-only memory, random access device, magnetic disk or optical disk, etc.

Landscapes

  • Physics & Mathematics (AREA)
  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Chemical & Material Sciences (AREA)
  • Analytical Chemistry (AREA)
  • Biochemistry (AREA)
  • General Health & Medical Sciences (AREA)
  • General Physics & Mathematics (AREA)
  • Immunology (AREA)
  • Pathology (AREA)
  • Analysing Materials By The Use Of Radiation (AREA)
  • Measurement Of Radiation (AREA)

Abstract

本申请实施例公开了一种检测方法和检测装置，所述方法包括：处理器从存储器获取与被测物对应的多个预设曝光时长，处理器根据探测器在每个预设曝光时长获取的放射线，得到每个预设曝光时长分别对应的被测物的放射线透射图；任一放射线透射图包括一个或者多个清晰的目标子区域；处理器识别每个放射线透射图中的目标子区域，根据识别出的所有目标子区域得到所述被测物的融合图像；处理器根据缺陷模型对融合图像进行识别，得到对被测物的检测结果。本申请实施例先用放射线照射被测物得到不同曝光时长对应的多个放射线透视图，然后根据缺陷模型对融合图像进行识别，这种方法提高了检测效率和检测精度。

Description

一种检测方法和检测装置 技术领域
本申请实施例涉及电子技术领域，尤其涉及一种检测方法和检测装置。
背景技术
为了确保电子产品正常发挥其功能，在电子产品使用之前，有必要对电子产品是否存在缺陷进行检测。传统的检测方法是人工使用放大镜对电子产品进行检测，由于电子产品具有结构紧凑和元器件设置集中的特点，使用传统的检测方法对电子产品进行检测时，检测结果依靠测试人员的视觉观察，效率低，精度没有保障。
发明内容
本申请实施例提供了一种检测方法和检测装置，能够提高检测效率和检测精度。
第一方面，本申请实施例提供了一种检测方法，应用于电子设备中，所述电子设备包括：放射线发射单元、探测器、处理器和存储器，被测物放置在所述放射线发射单元和所述探测器之间；所述方法包括如下步骤：
所述处理器从所述存储器获取与所述被测物对应的N个预设曝光时长，所述N为等于2或者大于2的整数，任一所述预设曝光时长对应的放射线透射图包括一个或者多个清晰的目标子区域；
所述处理器根据所述探测器在所述N个预设曝光时长中每个预设曝光时长获取的放射线，得到每个预设曝光时长分别对应的所述被测物的放射线透射图；
所述处理器识别每个所述放射线透射图中的目标子区域，根据识别出的目标子区域得到所述被测物的融合图像；
所述处理器根据缺陷模型对所述融合图像进行识别，得到对所述被测物的检测结果。
本申请实施例采用放射线照射被测物得到不同曝光时长对应的多个放射线透视图，识别每个放射线透射图中的目标子区域，根据识别出的目标子区域得到融合图像，然后根据缺陷模型对融合图像进行识别，得到对被测物的检测结果。由于融合图像由多个清晰的目标子区域得到，根据缺陷模型对融合图像进行识别，有利于提高检测效率和检测精度。
在本申请一些可能的实施方式中，所述电子设备还包括输出单元，所述方法还包括：
所述处理器向所述输出单元发送第一消息，所述第一消息携带所述检测结果；
所述输出单元根据接收到的所述第一消息输出所述检测结果。
在本申请一些可能的实施方式中，所述检测结果包括：正常、破损、毛边、刮伤、端子 歪、管脚缺失中的一个或者多个；
在本申请一些可能的实施方式中，输出单元可以包括显示屏，可以在显示屏中显示各种可能的检测结果的概率、或者各种检测结果的个数、或者结合表示被测物的图片将检测出缺陷的位置在图中标示出来、以及标注出检测到的缺陷。这样有利于直观地显示和观察被测物可能存在的缺陷。
在本申请一些可能的实施方式中，所述电子设备还包括采样单元，在所述处理器从所述存储器获取与所述被测物对应的N个预设曝光时长之前，所述方法还包括：
所述采样单元对所述被测物进行识别，并将识别结果发送给所述处理器；
所述处理器根据获取的所述识别结果从所述存储器获取与所述被测物对应的N个预设曝光时长。
其中，识别结果可以是被测物的类别标识，不同类别的被测物可能对应不同的曝光时长，可以在数据库中预先保存类别标识和其对应的曝光时长信息，处理器获取被测物的类别标识后查找数据库中保存的数据，得到被测物对应的N个曝光时长。
采用这种实施方式有利于对被测物进行自动检测，根据对被测物的识别结果处理器可以直接获取到与被测物对应的N个预设曝光时长。
在本申请一些可能的实施方式中，所述电子设备还包括传送单元；在所述采样单元对所述被测物进行识别之前，所述方法还包括：
所述传送单元每隔预设时长将不同的所述被测物移动到检测位置，所述检测位置位于所述放射线发射源和所述探测器之间。
采用该实施方式可以自动地将多个被测物逐一移动到检测位置进行检测，有利于提高检测效率。
在本申请一些可能的实施方式中，所述处理器根据缺陷模型对所述融合图像进行识别之前，所述方法还包括：
所述处理器根据预设的与所述被测物对应的缺陷训练样本进行缺陷数据训练和学习，得到所述被测物对应的缺陷模型。
需要说明的是，缺陷模型既可以由处理器根据预设的缺陷样本进行缺陷数据训练和学习获得，也可以从其他设备获取。
在本申请一些可能的实施方式中，所述处理器从所述存储器获取与所述被测物对应的N个预设曝光时长之前，所述方法还包括：
所述处理器根据所述被测物的结构和/或试验确定所述被测物对应的N个预设曝光时长，并将所述N个预设曝光时长保存到所述存储器中。
需要说明的是，不同物质对放射线的透射能力不同，根据被测物的结构或者通过试验，或者既通过参考被测物的结构又结合试验，确定多个曝光时长，有利于获取被测物在不同曝光时长时各部分的清晰图片，有利于提高检测的精度。
第二方面，本申请实施例提供了一种检测装置，所述检测装置包括：放射线发射单元、探测器、处理器和存储器，被测物放置在所述放射线发射单元和所述探测器之间；
所述放射线发射单元，用于发射放射线，所述放射线用于照射所述被测物；
所述探测器，用于接收经所述被测物透射的放射线；
所述处理器，用于从所述存储器获取与所述被测物对应的N个预设曝光时长，所述N为等于2或者大于2的整数，任一所述预设曝光时长对应的放射线透射图包括一个或者多个清晰的目标子区域；
所述处理器还用于，根据所述探测器在所述N个预设曝光时长中每个预设曝光时长获取的放射线，得到所述每个预设曝光时长分别对应的所述被测物的放射线透射图；
所述处理器还用于，识别每个所述放射线透射图中的目标子区域，根据识别出的目标子区域得到所述被测物的融合图像，并根据缺陷模型对所述融合图像进行识别，得到所述被测物的检测结果。
本申请实施例采用放射线照射被测物得到不同曝光时长对应的多个放射线透视图，然后识别每个放射线透射图中的目标子区域，根据识别出的目标子区域得到融合图像，然后根据缺陷模型对融合图像进行识别，得到对被测物的检测结果。由于融合图像由多个清晰的目标子区域得到，根据缺陷模型对融合图像进行识别，有利于提高检测效率和检测精度。
在本申请一些可能的实施方式中，所述电子设备还包括：输出单元；
所述处理器还用于，向所述输出单元发送第一消息，所述第一消息携带所述检测结果；
所述输出单元根据接收到的所述第一消息输出所述检测结果。
在本申请一些可能的实施方式中，所述检测结果包括：正常、破损、毛边、刮伤、端子歪、管脚缺失中的一个或者多个；
在本申请一些可能的实施方式中，输出单元可以包括显示屏，可以在显示屏中显示各种可能的检测结果的概率、或者各种检测结果的个数、或者结合表示被测物的图片将检测出缺陷的位置在图中标示出来、以及标注出检测到的缺陷。这样有利于直观地显示和观察被测物可能存在的缺陷。
在本申请一些可能的实施方式中，所述检测装置还包括：
采样单元，用于在所述处理器从所述存储器获取与所述被测物对应的N个预设曝光时长之前，对所述被测物进行识别，并将识别结果发送给所述处理器。
所述处理器，在从所述存储器获取与所述被测物对应的N个预设曝光时长方面，具体用于根据从所述采样单元获取的所述识别结果，从所述存储器获取与所述被测物对应的N个预设曝光时长。
在本申请一些可能的实施方式中，所述检测装置还包括：
传送单元，用于在所述采样单元对所述被测物进行识别之前，每隔预设时长将不同的所述被测物移动到检测位置，所述检测位置位于所述放射线发射源和所述探测器之间。
采用该实施方式可以自动地将多个被测物逐一移动到检测位置进行检测，有利于提高检测效率。
在本申请一些可能的实施方式中，所述处理器还用于，在根据缺陷模型和所述融合图像对所述被测物的缺陷进行识别之前，根据预设的与所述被测物对应的缺陷训练样本进行缺陷数据训练和学习，得到所述被测物对应的缺陷模型。
在本申请一些可能的实施方式中，所述处理器还用于，在从所述存储器获取与所述被测物对应的N个预设曝光时长之前，根据所述被测物的结构和/或试验确定所述被测物对应的N个预设曝光时长，并将所述N个预设曝光时长保持到所述存储器中。
需要说明的是，不同物质对放射线的透射能力不同，根据被测物的结构或者通过试验，或者既通过参考被测物的结构又结合试验，确定多个曝光时长，有利于获取被测物在不同时长下各部分的清晰图片，有利于提高检测的精度。
附图说明
为了更清楚地说明本发明实施例或现有技术中的技术方案，下面将对实施例或现有技术描述中所需要使用的附图作简单地介绍，显而易见地，下面描述中的附图仅仅是本发明的一些实施例，对于本领域普通技术人员来讲，在不付出创造性劳动的前提下，还可以根据这些附图获得其他的附图。
图1为本申请实施例提供的包括检测装置的电子设备应用场景示意图；
图2A为本申请实施例提供的一种检测方法的流程示意图；
图2B为本申请实施例提供的另一种检测方法的流程示意图；
图2C为本申请实施例提供的另一种检测方法的流程示意图；
图3A为本申请实施例提供的一种检测装置的结构示意图；
图3B为本申请实施例提供的另一种检测装置的结构示意图；
图3C为本申请实施例提供的另一种检测装置的结构示意图。
具体实施方式
为了使本技术领域的人员更好地理解本发明方案，下面将结合本发明实施例中的附图，对本发明实施例中的技术方案进行清楚、完整地描述，显然，所描述的实施例仅仅是本发明一部分实施例，而不是全部的实施例。基于本发明中的实施例，本领域普通技术人员在没有作出创造性劳动前提下所获得的所有其他实施例，都属于本发明保护的范围。
需要说明的是，在本发明实施例中使用的术语是仅仅出于描述特定实施例的目的，而非旨在限制本发明。在本发明实施例和所附权利要求书中所使用的单数形式的“一种”、“所述”和“该”也旨在包括多数形式，除非上下文清楚地表示其他含义。还应当理解，本文中使用 的术语“和/或”是指并包含一个或多个相关联的列出项目的任何或所有可能组合。
需要说明的是，下面详细描述本发明的实施例，所述实施例的示例在附图中示出，其中自始至终相同或类似的标号表示相同或类似的元件或具有相同或类似功能的元件。下面通过参考附图描述的实施例是示例性的，仅用于解释本发明，而不能解释为对本发明的限制。
下文的公开提供了许多不同的实施例或例子用来实现本发明的不同结构。为了简化本发明的公开，下文中对特定例子的部件和设置进行描述。当然，它们仅仅为示例，并且目的不在于限制本发明。此外，本发明可以在不同例子中重复参考数字和/或字母。这种重复是为了简化和清楚的目的，其本身不指示所讨论各种实施例和/或设置之间的关系。
连接器的形式和结构千变万化，随着应用对象、频率、功率、应用环境等的不同，有各种不同形式的连接器，但是无论是哪种连接器，在其正常工作时，都要保证电流顺畅连续和可靠地流通，连接器中的针pin是连接器的重要结构，主要用于导电和信号传输，连接器在生产过程中，pin脚在组装时，由于气缸运动存在受力不均的情况，导致可能出现pin脚出现倾斜，影响产品的最终质量，因此有必要对连接器的pin脚进行检测，传统的检测方法是检测人员使用放大镜检测连接器是否合格，这种方法完全依赖检测人员的视觉直觉，准确率低，精度没有保障。本申请实施例使用放射线照射被测物，利用不同曝光时长对应的放射线透射图，得到多个目标子区域，根据目标子区域得到被测物的融合图像，然后根据缺陷模型对融合图像进行识别，得到被测物的检测结果。由于融合图像由多个清晰的目标子区域得到，根据缺陷模型对融合图像进行识别，有利于提高检测效率和检测精度。
请参阅图1，图1为本申请实施例提供了包括检测装置的电子设备应用场景示意图。如图1所示，电子装置可以包括：放射线发射单元101、探测器103、存储器104和处理器105，被测物102放置在放射线发射单元101和探测器103之间；在对被测物102进行检测时，放射线发射单元101发出放射线，照射被测物102，探测器103根据穿过被测物102的放射线生成与被测物102对应的图像，处理器105从存储器104获取与被测物102对应的N个预设曝光时长，所述N为等于2或者大于2的整数，任一预设曝光时长对应的放射线透射图包括一个或者多个清晰的目标子区域；处理器105根据探测器103在N个预设曝光时长中每个预设曝光时长获取的放射线，得到每个预设曝光时长分别对应的被测物102的放射线透射图；处理器105识别每个放射线透射图中的目标子区域，根据识别出的目标子区域得到被测物102的融合图像，即包括多个目标子区域的图像；处理器105根据缺陷模型对融合图像进行识别，得到对被测物102的检测结果。由于融合图像由多个清晰的目标子区域得到，根据缺陷模型对融合图像进行识别，有利于提高检测效率和检测精度。
参见图2A，图2A为本申请实施例提供的一种检测方法的流程示意图，检测方法应用于电子设备中，电子设备包括：放射线发射单元、探测器、处理器和存储器，被测物放置在所述放射线发射单元和所述探测器之间；所述方法包括步骤201-204，具体如下：
201、处理器从所述存储器获取与所述被测物对应的N个预设曝光时长，所述N为等于2或者大于2的整数，任一所述预设曝光时长对应的放射线透射图包括一个或者多个清晰的目标子区域。
在一些可能的实施方式中，处理器可以根据被测物的结构确定被测物对应的N个预设曝光时长，并将所述N个预设曝光时长保存到存储器中。比如，若被测物包括3个不同的材质组成的结构G1、G2和G3，若这三个不同的结构在不同曝光时长时对应的清晰度不同，可以设定N＝3，根据实验确定不同的结构在曝光多久时可以得到清晰的图像，举例来说，若结构G1在曝光时长为2秒时处理器根据探测器接收到的放射线可以得到结构G1的清晰图像，则结构G1对应的预设曝光时长为2秒。若结构G2在曝光时长为5秒时处理器根据探测器接收到的放射线可以得到结构G2的清晰图像，则结构G2对应的预设曝光时长为5秒。若结构G3在曝光时长为9秒时处理器根据探测器接收到的放射线可以得到结构G3的清晰图像，结构G3对应的预设曝光时长为9秒。因此，可以预先设置与被测物对应的3个预设曝光时长分别为2秒、5秒和9秒。
可以理解的是，在一些可能的实施方式中，被测物不同位置的材质是相同的，因为形状结构等存在差异，比如薄厚不同，也会使得被测物的不同位置在不同曝光时长时得到的图像清晰度不同，根据不同曝光时长得到的图像可以对不同位置进行检测。可以理解的，被测物预设时长的确定可以通过试验确定，确定后将预设时长保存在存储器中，在检测某个被测物时，从存储器中取出与该被测物对应的预设时长。
202、所述处理器根据所述探测器在所述N个预设曝光时长中每个预设曝光时长获取的放射线，得到每个预设曝光时长分别对应的所述被测物的放射线透射图。
处理器根据探测器在不同曝光时长获取的放射线，可以得到在不同曝光时长下与被测物对应的放射线透射图。以图1为例，被测物102包括三个不同的区域：区域A1、区域A2和区域An，其中，区域A1在曝光时长为T1时对应的放射线透射图中最清楚，区域A2在曝光时长为T2时对应的放射线透射图中最清楚，区域An在曝光时长为Tn时对应的放射线透射图中最清楚。则处理器在被测物102曝光时长分别为T1、T2、Tn时根据探测器接收到的放射线，生成被测物的放射线透射图p1、p2和p3。
203、处理器识别每个所述放射线透射图中的目标子区域，根据识别出的目标子区域得到所述被测物的融合图像。
以图1中被测物102为例，处理器105在获取被测物102在曝光时长分别为T1、T2、Tn时根据探测器接收到的放射线生成被测物的放射线透射图p1、p2和p3之后，处理器识别各透视图中清晰的区域即各幅图片的目标子区域，比如从图p1中识别目标子区域A1，从图p2中识别目标子区域A2，从图p3中识别目标子区域An。然后处理器将提取出来的各目标子区域进行处理得到包括A1、A2和A3的融合图像p。
需要说明的是，当被测物是连接器时，连接器上不同结构件的材料不同，使得连接器上 不同结构件的成像质量受曝光时间因素影响程度不同，可以表现在亮度、对比度、边缘信息等方面，可以将以上参数良好的区域成为有利区域，放射线可以是X射线，X射线在不同曝光时间下照射被测物，得到有利区域不同的图像，将有利区域交集最小且并集最大的图像进行融合，得到各部分都清晰的图像作为融合图像，用于进一步检测确定连接器是否存在缺陷。
204、所述处理器根据缺陷模型对所述融合图像进行识别，得到对所述被测物的检测结果。
在一些可能的实施方式中，处理器根据预设的与被测物对应的缺陷训练样本进行缺陷数据训练和学习，得到被测物对应的缺陷模型。以图1为例，若经过识别确定A1对应的区域不存在缺陷，A2和A3对应的区域存在缺陷时，则可以将包括存在缺陷的区域对应图像显示出来作为检测结果。可以理解的，也可以进一步的通过语音或者图标或者动态提醒图片提醒有缺陷存在。
在一些可能的实施方式中，处理器可以利用深度学习神经网络自动对被测物的缺陷进行识别。在对被测物进行检测之前，预先通过大量的训练样本，对被测物不同的缺陷进行识别，比如，按照不同的缺陷类型分别采集100幅以上的图像作为训练样本，对于能够直接反应缺陷的样本进行分类训练，得到训练后的被测物缺陷图像，在进行检测时，图像融合后的图像与预先得到的被测物的缺陷图像进行比较，识别出可能存在的缺陷，识别结果可能包括：正常、破损、毛边、刮伤、端子歪、针脚缺失等。
本申请实施例采用放射线照射被测物得到不同曝光时长对应的多个放射线透视图，识别每个放射线透射图中的目标子区域，根据识别出的目标子区域得到融合图像，然后根据缺陷模型对融合图像进行识别，得到对被测物的检测结果。由于融合图像由多个清晰的目标子区域得到，根据缺陷模型对融合图像进行识别，有利于提高检测效率和检测精度。
参见图2B，图2B为本申请另一实施例提供的一种检测方法的流程示意图，检测方法应用于电子设备中，电子设备包括：放射线发射单元、探测器、处理器和存储器，被测物放置在所述放射线发射单元和所述探测器之间；所述方法包括步骤211-215，具体如下：
211、采样单元对被测物进行识别，并将识别结果发送给处理器。
比如当被测物放置在放射线发射单元和探测器之间之后，采样单元开始对被测物进行检测，以被测物为连接器为例，当连接器放置在放射线发射单元和探测器之间之后，采样单元对被测物进行识别，识别出被测物为连接器。
212、所述处理器根据获取的所述识别结果从所述存储器获取与所述被测物对应的N个预设曝光时长。所述N为等于2或者大于2的整数，任一所述预设曝光时长对应的放射线透射图包括一个或者多个清晰的目标子区域。
需要说明的是，在存储器中保存了至少一个检测对象对应的预设曝光时长，比如在存储器中预先保存了连接器对应的N各预设曝光时长，比如，N可以为3，预设曝光时长可以为2秒、5秒和9秒。
在一些可能的实施方式中，处理器可以根据被测物的结构确定被测物对应的N个预设曝光时长，并将所述N个预设曝光时长保存到存储器中。比如，若被测物包括3个不同的材质组成的结构G1、G2和G3，若这三个不同的结构在不同曝光时长时对应的清晰度不同，可以设定N＝3，根据实验确定不同的结构在曝光多久时可以得到清晰的图像，举例来说，若结构G1在曝光时长为2秒时处理器根据探测器接收到的放射线可以得到结构G1的清晰图像，则结构G1对应的预设曝光时长为2秒。若结构G2在曝光时长为5秒时处理器根据探测器接收到的放射线可以得到结构G2的清晰图像，则结构G2对应的预设曝光时长为5秒。若结构G3在曝光时长为9秒时处理器根据探测器接收到的放射线可以得到结构G3的清晰图像，结构G3对应的预设曝光时长为9秒。因此，可以预先设置与被测物对应的3个预设曝光时长分别为2秒、5秒和9秒。
可以理解的是，在一些可能的实施方式中，被测物不同位置的材质是相同的，因为形状结构等存在差异，比如薄厚不同，也会使得被测物的不同位置在不同曝光时长时得到的图像清晰度不同，根据不同曝光时长得到的图像可以对不同位置进行检测。可以理解的，被测物预设时长的确定可以通过试验确定，确定后将预设时长保存在存储器中，在检测某个被测物时，从存储器中取出与该被测物对应的预设时长。
213、所述处理器根据所述探测器在所述N个预设曝光时长中每个预设曝光时长获取的放射线，得到每个预设曝光时长分别对应的所述被测物的放射线透射图。
处理器根据探测器在不同曝光时长获取的放射线，可以得到在不同曝光时长下与被测物对应的放射线透射图。以图1为例，被测物102包括三个不同的区域：区域A1、区域A2和区域An，其中，区域A1在曝光时长为T1时对应的放射线透射图中最清楚，区域A2在曝光时长为T2时对应的放射线透射图中最清楚，区域An在曝光时长为Tn时对应的放射线透射图中最清楚。则处理器在被测物102曝光时长分别为T1、T2、Tn时根据探测器接收到的放射线，生成被测物的放射线透射图p1、p2和p3。
214、处理器识别每个所述放射线透射图中的目标子区域，根据识别出的目标子区域得到所述被测物的融合图像。
以图1中被测物102为例，处理器105在获取被测物102在曝光时长分别为T1、T2、Tn时根据探测器接收到的放射线生成被测物的放射线透射图p1、p2和p3之后，处理器识别各透视图中清晰的区域即各幅图片的目标子区域，比如从图p1中识别目标子区域A1，从图p2中识别目标子区域A2，从图p3中识别目标子区域An。然后处理器将提取出来的各目标子区域进行处理得到包括A1、A2和A3的融合图像p。
需要说明的是，当被测物是连接器时，连接器上不同结构件的材料不同，使得连接器上不同结构件的成像质量受曝光时间因素影响程度不同，可以表现在亮度、对比度、边缘信息等方面，可以将以上参数良好的区域成为有利区域，放射线可以是X射线，X射线在不同曝光时间下照射被测物，得到有利区域不同的图像，将有利区域交集最小且并集最大的图像进 行融合，得到各部分都清晰的图像作为融合图像，用于进一步检测确定连接器是否存在缺陷。215、所述处理器根据缺陷模型对所述融合图像进行识别，得到对所述被测物的检测结果。
在一些可能的实施方式中，处理器根据预设的与被测物对应的缺陷训练样本进行缺陷数据训练和学习，得到被测物对应的缺陷模型。以图1为例，若经过识别确定A1对应的区域不存在缺陷，A2和A3对应的区域存在缺陷时，则可以将包括存在缺陷的区域对应图像显示出来作为检测结果。可以理解的，也可以进一步的通过语音或者图标或者动态提醒图片提醒有缺陷存在。
在一些可能的实施方式中，处理器可以利用深度学习神经网络自动对被测物的缺陷进行识别。在对被测物进行检测之前，预先通过大量的训练样本，对被测物不同的缺陷进行识别，比如，按照不同的缺陷类型分别采集100幅以上的图像作为训练样本，对于能够直接反应缺陷的样本进行分类训练，得到训练后的被测物缺陷图像，在进行检测时，图像融合后的图像与预先得到的被测物的缺陷图像进行比较，识别出可能存在的缺陷，识别结果可能包括：正常、破损、毛边、刮伤、端子歪、针脚缺失等。
参见图2C，图2C为本申请另一实施例提供的一种检测方法的流程示意图，该检测方法应用于电子设备中，电子设备包括：放射线发射单元、探测器、处理器和存储器，被测物放置在所述放射线发射单元和所述探测器之间；所述方法包括步骤221-226，具体如下：
221、传送单元每隔预设时长将不同的被测物传送到检测位置，所述检测位置位于所述放射线发射源和所述探测器之间。
在一些可能的实施例中，传送带传送被测物的间隔时长由对对被测物的检测速度来确定，可以设置当前被测物的检测结果出来之后，下一个被测物才经传送带传送到放射线发射单元和探测器之间。比如，若一个被测物的检测周期为15秒，则传送带每隔15秒将另外一个被测物传送到检测位置。可以理解的是，不同的被测物对应传送带的传送速度也不同，比如若第一类被测物的检测时长为15秒，第二类被测物的检测时长为30秒，则当对第一类被测物进行检测时，传送带每隔15秒将下一个第一类被测物传送到被测位置。当对第二类被测物进行检测时，传送带每隔30秒将下一个第二类被测物传送到被测位置。可以理解的，传送带传送速度也可以采用检测时长最长的被测物对应的检测时间作为传送带传送相邻两个被测物的间隔时长。
222、采样单元对被测物进行识别，并将识别结果发送给处理器。
比如当被测物放置在放射线发射单元和探测器之间之后，采样单元开始对被测物进行检测，以被测物为连接器为例，当连接器放置在放射线发射单元和探测器之间之后，采样单元对被测物进行识别，识别出被测物为连接器。
223、所述处理器根据获取的所述识别结果从所述存储器获取与所述被测物对应的N个预设曝光时长。所述N为等于2或者大于2的整数，任一所述预设曝光时长对应的放射线透 射图包括一个或者多个清晰的目标子区域。
需要说明的是，在存储器中保存了至少一个检测对象对应的预设曝光时长，比如在存储器中预先保存了连接器对应的N各预设曝光时长，比如，N可以为3，预设曝光时长可以为2秒、5秒和9秒。
在一些可能的实施方式中，处理器可以根据被测物的结构确定被测物对应的N个预设曝光时长，并将所述N个预设曝光时长保存到存储器中。比如，若被测物包括3个不同的材质组成的结构G1、G2和G3，若这三个不同的结构在不同曝光时长时对应的清晰度不同，可以设定N＝3，根据实验确定不同的结构在曝光多久时可以得到清晰的图像，举例来说，若结构G1在曝光时长为2秒时处理器根据探测器接收到的放射线可以得到结构G1的清晰图像，则结构G1对应的预设曝光时长为2秒。若结构G2在曝光时长为5秒时处理器根据探测器接收到的放射线可以得到结构G2的清晰图像，则结构G2对应的预设曝光时长为5秒。若结构G3在曝光时长为9秒时处理器根据探测器接收到的放射线可以得到结构G3的清晰图像，结构G3对应的预设曝光时长为9秒。因此，可以预先设置与被测物对应的3个预设曝光时长分别为2秒、5秒和9秒。
可以理解的是，在一些可能的实施方式中，被测物不同位置的材质是相同的，因为形状结构等存在差异，比如薄厚不同，也会使得被测物的不同位置在不同曝光时长时得到的图像清晰度不同，根据不同曝光时长得到的图像可以对不同位置进行检测。可以理解的，被测物预设时长的确定可以通过试验确定，确定后将预设时长保存在存储器中，在检测某个被测物时，从存储器中取出与该被测物对应的预设时长。
224、所述处理器根据所述探测器在所述N个预设曝光时长中每个预设曝光时长获取的放射线，得到每个预设曝光时长分别对应的所述被测物的放射线透射图。
处理器根据探测器在不同曝光时长获取的放射线，可以得到在不同曝光时长下与被测物对应的放射线透射图。以图1为例，被测物102包括三个不同的区域：区域A1、区域A2和区域An，其中，区域A1在曝光时长为T1时对应的放射线透射图中最清楚，区域A2在曝光时长为T2时对应的放射线透射图中最清楚，区域An在曝光时长为Tn时对应的放射线透射图中最清楚。则处理器在被测物102曝光时长分别为T1、T2、Tn时根据探测器接收到的放射线，生成被测物的放射线透射图p1、p2和p3。
225、处理器识别每个所述放射线透射图中的目标子区域，根据识别出的目标子区域得到所述被测物的融合图像。
以图1中被测物102为例，处理器105在获取被测物102在曝光时长分别为T1、T2、Tn时根据探测器接收到的放射线生成被测物的放射线透射图p1、p2和p3之后，处理器识别各透视图中清晰的区域即各幅图片的目标子区域，比如从图p1中识别目标子区域A1，从图p2中识别目标子区域A2，从图p3中识别目标子区域An。然后处理器将提取出来的各目标子区域进行处理得到包括A1、A2和A3的融合图像p。
需要说明的是，当被测物是连接器时，连接器上不同结构件的材料不同，使得连接器上不同结构件的成像质量受曝光时间因素影响程度不同，可以表现在亮度、对比度、边缘信息等方面，可以将以上参数良好的区域成为有利区域，放射线可以是X射线，X射线在不同曝光时间下照射被测物，得到有利区域不同的图像，将有利区域交集最小且并集最大的图像进行融合，得到各部分都清晰的图像作为融合图像，用于进一步检测确定连接器是否存在缺陷。
226、所述处理器根据缺陷模型对所述融合图像进行识别，得到对所述被测物的检测结果。
采用该实施方式传送单元可以自动地将多个被测物逐一移动到检测位置进行检测，有利于提高检测效率。
在本申请一些可能的实施方式中，所述电子设备还包括输出单元，所述方法还包括：
所述处理器向所述输出单元发送第一消息，所述第一消息携带所述检测结果；
所述输出单元根据接收到的所述第一消息输出所述检测结果。
在本申请一些可能的实施方式中，输出单元可以包括显示屏，可以在显示屏中显示各种可能的检测结果的概率、或者各种检测结果的个数、或者结合表示被测物的图片将检测出缺陷的位置在图中标示出来、以及标注出检测到的缺陷。这样有利于直观地显示和观察被测物可能存在的缺陷。
上述主要从方法侧执行过程的角度对本申请实施例的方案进行了介绍。可以理解的是，终端为了实现上述功能，其包含了执行各个功能相应的硬件结构和/或软件模块。本领域技术人员应该很容易意识到，结合本文中所提供的实施例描述的各示例的单元及算法步骤，本申请能够以硬件或硬件和计算机软件的结合形式来实现。某个功能究竟以硬件还是计算机软件驱动硬件的方式来执行，取决于技术方案的特定应用和设计约束条件。专业技术人员可以对每个特定的应用使用不同方法来实现所描述的功能，但是这种实现不应认为超出本申请的范围。
本申请实施例可以根据上述方法示例对终端进行功能单元的划分，例如，可以对应各个功能划分各个功能单元，也可以将两个或两个以上的功能集成在一个处理单元中。上述集成的单元既可以采用硬件的形式实现，也可以采用软件功能单元的形式实现。需要说明的是，本申请实施例中对单元的划分是示意性的，仅仅为一种逻辑功能划分，实际实现时可以有另外的划分方式。
请参阅图3A，图3A为本申请实施例提供的一种检测装置的结构示意图。如图3A所示，检测装置310，包括：放射线发射单元311、探测器312、处理器313和存储器314，被测物300放置在放射线发射单元311和探测器312之间；
放射线发射单元311，用于发射放射线，所述放射线用于照射所述被测物。
探测器312，用于接收经所述被测物透射的放射线。
处理器313，用于从所述存储器获取与所述被测物对应的N个预设曝光时长，所述N为 等于2或者大于2的整数，任一所述预设曝光时长对应的放射线透射图包括一个或者多个清晰的目标子区域。
处理器313还用于，根据所述探测器在所述N个预设曝光时长中每个预设曝光时长获取的放射线，得到所述每个预设曝光时长分别对应的所述被测物的放射线透射图。
在一些可能的实施方式中，处理器可以根据被测物的结构确定被测物对应的N个预设曝光时长，并将所述N个预设曝光时长保存到存储器中。比如，若被测物包括3个不同的材质组成的结构G1、G2和G3，若这三个不同的结构在不同曝光时长时对应的清晰度不同，可以设定N＝3，根据实验确定不同的结构在曝光多久时可以得到清晰的图像，举例来说，若结构G1在曝光时长为2秒时处理器根据探测器接收到的放射线可以得到结构G1的清晰图像，则结构G1对应的预设曝光时长为2秒。若结构G2在曝光时长为5秒时处理器根据探测器接收到的放射线可以得到结构G2的清晰图像，则结构G2对应的预设曝光时长为5秒。若结构G3在曝光时长为9秒时处理器根据探测器接收到的放射线可以得到结构G3的清晰图像，结构G3对应的预设曝光时长为9秒。因此，可以预先设置与被测物对应的3个预设曝光时长分别为2秒、5秒和9秒。
可以理解的是，在一些可能的实施方式中，被测物不同位置的材质是相同的，因为形状结构等存在差异，比如薄厚不同，也会使得被测物的不同位置在不同曝光时长时得到的图像清晰度不同，根据不同曝光时长得到的图像可以对不同位置进行检测。可以理解的，被测物预设时长的确定可以通过试验确定，确定后将预设时长保存在存储器中，在检测某个被测物时，从存储器中取出与该被测物对应的预设时长。
处理器根据探测器在不同曝光时长获取的放射线，可以得到在不同曝光时长下与被测物对应的放射线透射图。以图1为例，被测物102包括三个不同的区域：区域A1、区域A2和区域An，其中，区域A1在曝光时长为T1时对应的放射线透射图中最清楚，区域A2在曝光时长为T2时对应的放射线透射图中最清楚，区域An在曝光时长为Tn时对应的放射线透射图中最清楚。则处理器在被测物102曝光时长分别为T1、T2、Tn时根据探测器接收到的放射线，生成被测物的放射线透射图p1、p2和p3。
处理器313还用于，识别每个所述放射线透射图中的目标子区域，根据识别出的目标子区域得到所述被测物的融合图像，并根据缺陷模型对所述融合图像进行识别，得到所述被测物的检测结果。
以图1中被测物102为例，处理器105在获取被测物102在曝光时长分别为T1、T2、Tn时根据探测器接收到的放射线生成被测物的放射线透射图p1、p2和p3之后，处理器识别各透视图中清晰的区域即各幅图片的目标子区域，比如从图p1中识别目标子区域A1，从图p2中识别目标子区域A2，从图p3中识别目标子区域An。然后处理器将提取出来的各目标子区域进行处理得到包括A1、A2和A3的融合图像p。
需要说明的是，当被测物是连接器时，连接器上不同结构件的材料不同，使得连接器上 不同结构件的成像质量受曝光时间因素影响程度不同，可以表现在亮度、对比度、边缘信息等方面，可以将以上参数良好的区域成为有利区域，放射线可以是X射线，X射线在不同曝光时间下照射被测物，得到有利区域不同的图像，将有利区域交集最小且并集最大的图像进行融合，得到各部分都清晰的图像作为融合图像，用于进一步检测确定连接器是否存在缺陷。
在一些可能的实施方式中，处理器根据预设的与被测物对应的缺陷训练样本进行缺陷数据训练和学习，得到被测物对应的缺陷模型。以图1为例，若经过识别确定A1对应的区域不存在缺陷，A2和A3对应的区域存在缺陷时，则可以将包括存在缺陷的区域对应图像显示出来作为检测结果。可以理解的，也可以进一步的通过语音或者图标或者动态提醒图片提醒有缺陷存在。
本申请实施例采用放射线照射被测物得到不同曝光时长对应的多个放射线透视图，识别每个放射线透射图中的目标子区域，根据识别出的目标子区域得到融合图像，然后根据缺陷模型对融合图像进行识别，得到对被测物的检测结果。由于融合图像由多个清晰的目标子区域得到，根据缺陷模型对融合图像进行识别，有利于提高检测效率和检测精度。
请参阅图3B，图3B为本申请实施例提供的另一检测装置的结构示意图。如图3B所示，该实施例相对于图3A所示实施例，检测装置310还包括采样单元315。
采样单元315，用于在处理器313从存储器314获取与被测物300对应的N个预设曝光时长之前，对被测物300进行识别，并将识别结果发送给处理器313。
比如当被测物放置在放射线发射单元和探测器之间之后，采样单元开始对被测物进行检测，以被测物为连接器为例，当连接器放置在放射线发射单元和探测器之间之后，采样单元对被测物进行识别，识别出被测物为连接器。
处理器313，在从所述存储器获取与所述被测物对应的N个预设曝光时长方面，具体用于根据从所述采样单元获取的所述识别结果从所述存储器获取与所述被测物对应的N个预设曝光时长。
需要说明的是，在存储器中保存了至少一个检测对象对应的预设曝光时长，比如在存储器中预先保存了连接器对应的N各预设曝光时长，比如，N可以为3，预设曝光时长可以为2秒、5秒和9秒。
在一些可能的实施方式中，处理器可以根据被测物的结构确定被测物对应的N个预设曝光时长，并将所述N个预设曝光时长保存到存储器中。比如，若被测物包括3个不同的材质组成的结构G1、G2和G3，若这三个不同的结构在不同曝光时长时对应的清晰度不同，可以设定N＝3，根据实验确定不同的结构在曝光多久时可以得到清晰的图像，举例来说，若结构G1在曝光时长为2秒时处理器根据探测器接收到的放射线可以得到结构G1的清晰图像，则结构G1对应的预设曝光时长为2秒。若结构G2在曝光时长为5秒时处理器根据探测器接收到的放射线可以得到结构G2的清晰图像，则结构G2对应的预设曝光时长为5秒。若结构 G3在曝光时长为9秒时处理器根据探测器接收到的放射线可以得到结构G3的清晰图像，结构G3对应的预设曝光时长为9秒。因此，可以预先设置与被测物对应的3个预设曝光时长分别为2秒、5秒和9秒。
可以理解的是，在一些可能的实施方式中，被测物不同位置的材质是相同的，因为形状结构等存在差异，比如薄厚不同，也会使得被测物的不同位置在不同曝光时长时得到的图像清晰度不同，根据不同曝光时长得到的图像可以对不同位置进行检测。可以理解的，被测物预设时长的确定可以通过试验确定，确定后将预设时长保存在存储器中，在检测某个被测物时，从存储器中取出与该被测物对应的预设时长。
采用这种实施方式有利于对被测物进行自动检测，根据对被测物的识别结果处理器可以直接获取到与被测物对应的N个预设曝光时长。
请参阅图3C，图3C为本申请实施例提供的另一检测装置的结构示意图。如图3C所示，该实施例相对于图3B所示实施例，检测装置310还包括传送单元316。
传送单元316用于在采样单元315对被测物300进行识别之前，每隔预设时长将不同的被测物300传送到检测位置，检测位置位于放射线发射单元311和探测器312之间。
在一些可能的实施例中，传送带传送被测物的间隔时长由对对被测物的检测速度来确定，可以设置当前被测物的检测结果出来之后，下一个被测物才经传送带传送到放射线发射单元和探测器之间。比如，若一个被测物的检测周期为15秒，则传送带每隔15秒将另外一个被测物传送到检测位置。可以理解的是，不同的被测物对应传送带的传送速度也不同，比如若第一类被测物的检测时长为15秒，第二类被测物的检测时长为30秒，则当对第一类被测物进行检测时，传送带每隔15秒将下一个第一类被测物传送到被测位置。当对第二类被测物进行检测时，传送带每隔30秒将下一个第二类被测物传送到被测位置。可以理解的，传送带传送速度也可以采用检测时长最长的被测物对应的检测周期作为传送带传送相邻两个被测物的间隔时长。
采用该实施方式传送单元可以自动地将多个被测物逐一移动到检测位置进行检测，有利于提高检测效率。
在本申请一些可能的实施方式中，检测装置还可以包括：输出单元；所述处理器还用于，向所述输出单元发送第一消息，所述第一消息携带所述检测结果；所述输出单元根据接收到的所述第一消息输出所述检测结果。所述检测结果包括：正常、破损、毛边、刮伤、端子歪、管脚缺失中的一个或者多个；输出单元可以包括显示屏，可以在显示屏中显示各种可能的检测结果的概率、或者各种检测结果的个数、或者结合表示被测物的图片将检测出缺陷的位置在图中标示出来、以及标注出检测到的缺陷。这样有利于直观地显示和观察被测物可能存在的缺陷。
本申请实施例还提供了包括检测装置的电子设备，检测装置可以为前面任一实施例所述的检测装置。
本申请实施例还提供一种计算机存储介质，其中，该计算机存储介质存储用于电子数据交换的计算机程序，该计算机程序使得计算机执行如上述方法实施例中记载的任何一种检测方法的部分或全部步骤。
本申请实施例还提供一种计算机程序产品，所述计算机程序产品包括存储了计算机程序的非瞬时性计算机可读存储介质，该计算机程序使得计算机执行如上述方法实施例中记载的任何一种检测方法的部分或全部步骤。
需要说明的是，对于前述的各方法实施例，为了简单描述，故将其都表述为一系列的动作组合，但是本领域技术人员应该知悉，本申请并不受所描述的动作顺序的限制，因为依据本申请，某些步骤可以采用其他顺序或者同时进行。其次，本领域技术人员也应该知悉，说明书中所描述的实施例均属于优选实施例，所涉及的动作和模块并不一定是本申请所必须的。
在上述实施例中，对各个实施例的描述都各有侧重，某个实施例中没有详述的部分，可以参见其他实施例的相关描述。
在本申请所提供的几个实施例中，应该理解到，所揭露的装置，可通过其它的方式实现。例如，以上所描述的装置实施例仅仅是示意性的，例如所述单元的划分，仅仅为一种逻辑功能划分，实际实现时可以有另外的划分方式，例如多个单元或组件可以结合或者可以集成到另一个系统，或一些特征可以忽略，或不执行。另一点，所显示或讨论的相互之间的耦合或直接耦合或通信连接可以是通过一些接口，装置或单元的间接耦合或通信连接，可以是电性或其它的形式。
本领域普通技术人员可以理解上述实施例的各种方法中的全部或部分步骤是可以通过程序来指令相关的硬件来完成，该程序可以存储于一计算机可读存储器中，存储器可以包括：闪存盘、只读存储器、随机存取器、磁盘或光盘等。
以上所揭露的仅为本发明一种较佳实施例而已，当然不能以此来限定本发明之权利范围，本领域普通技术人员可以理解实现上述实施例的全部或部分流程，并依本发明权利要求所作的等同变化，仍属于发明所涵盖的范围。

Claims (10)

  1. 一种检测方法，其特征在于，应用于电子设备中，所述电子设备包括：放射线发射单元、探测器、处理器和存储器，被测物放置在所述放射线发射单元和所述探测器之间；所述方法包括如下步骤：
    所述处理器从所述存储器获取与所述被测物对应的N个预设曝光时长，所述N为等于2或者大于2的整数，任一所述预设曝光时长对应的放射线透射图包括一个或者多个清晰的目标子区域；
    所述处理器根据所述探测器在所述N个预设曝光时长中每个预设曝光时长获取的放射线，得到每个预设曝光时长分别对应的所述被测物的放射线透射图；
    所述处理器识别每个所述放射线透射图中的目标子区域，根据识别出的目标子区域得到所述被测物的融合图像；
    所述处理器根据缺陷模型对所述融合图像进行识别，得到对所述被测物的检测结果。
  2. 根据权利要求1所述的方法，其特征在于，所述电子设备还包括采样单元，在所述处理器从所述存储器获取与所述被测物对应的N个预设曝光时长之前，所述方法还包括：
    所述采样单元对所述被测物进行识别，并将识别结果发送给所述处理器；
    所述处理器根据获取的所述识别结果从所述存储器获取与所述被测物对应的N个预设曝光时长。
  3. 根据权利要求2所述的方法，其特征在于，所述电子设备还包括传送单元；在所述采样单元对所述被测物进行识别之前，所述方法还包括：
    所述传送单元每隔预设时长将不同的所述被测物传送到检测位置，所述检测位置位于所述放射线发射源和所述探测器之间。
  4. 根据权利要求1所述的方法，其特征在于，所述处理器根据缺陷模型对所述融合图像进行识别之前，所述方法还包括：
    所述处理器根据预设的与所述被测物对应的缺陷训练样本进行缺陷数据训练和学习，得到所述被测物对应的缺陷模型。
  5. 根据权利要求1至4任一项所述的方法，其特征在于，所述处理器从所述存储器获取与所述被测物对应的N个预设曝光时长之前，所述方法还包括：
    所述处理器根据所述被测物的结构和/或试验确定所述被测物对应的N个预设曝光时长，并将所述N个预设曝光时长保存到所述存储器中。
  6. 一种检测装置，其特征在于，包括：放射线发射单元、探测器、处理器和存储器，被测物放置在所述放射线发射单元和所述探测器之间；
    所述放射线发射单元，用于发射放射线，所述放射线用于照射所述被测物；
    所述探测器，用于接收经所述被测物透射的放射线；
    所述处理器，用于从所述存储器获取与所述被测物对应的N个预设曝光时长，所述N为等于2或者大于2的整数，任一所述预设曝光时长对应的放射线透射图包括一个或者多个清晰的目标子区域；
    所述处理器还用于，根据所述探测器在所述N个预设曝光时长中每个预设曝光时长获取的放射线，得到所述每个预设曝光时长分别对应的所述被测物的放射线透射图；
    所述处理器还用于，识别每个所述放射线透射图中的目标子区域，根据识别出的目标子区域得到所述被测物的融合图像，并根据缺陷模型对所述融合图像进行识别，得到所述被测物的检测结果。
  7. 根据权利要求6所述的检测装置，其特征在于，所述检测装置还包括：
    采样单元，用于在所述处理器从所述存储器获取与所述被测物对应的N个预设曝光时长之前，对所述被测物进行识别，并将识别结果发送给所述处理器；
    所述处理器，在从所述存储器获取与所述被测物对应的N个预设曝光时长方面，具体用于根据从所述采样单元获取的所述识别结果从所述存储器获取与所述被测物对应的N个预设曝光时长。
  8. 根据权利要求7所述的检测装置，其特征在于，所述检测装置还包括：
    传送单元，用于在所述采样单元对所述被测物进行识别之前，每隔预设时长将不同的所述被测物传送到检测位置，所述检测位置位于所述放射线发射源和所述探测器之间。
  9. 根据权利要求6所述的检测装置，其特征在于，
    所述处理器还用于，在根据缺陷模型对所述融合图像进行识别之前，根据预设的与所述被测物对应的缺陷训练样本进行缺陷数据训练和学习，得到所述被测物对应的缺陷模型。
  10. 根据权利要求6至9任一项所述的检测装置，其特征在于，
    所述处理器还用于，在从所述存储器获取与所述被测物对应的N个预设曝光时长之前，根据所述被测物的结构和/或试验确定所述被测物对应的N个预设曝光时长，并将所述N个预设曝光时长保存到所述存储器中。
PCT/CN2018/112272 2018-10-27 2018-10-27 一种检测方法和检测装置 WO2020082396A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
CN201880068619.2A CN111356914B (zh) 2018-10-27 2018-10-27 一种检测方法和检测装置
PCT/CN2018/112272 WO2020082396A1 (zh) 2018-10-27 2018-10-27 一种检测方法和检测装置

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2018/112272 WO2020082396A1 (zh) 2018-10-27 2018-10-27 一种检测方法和检测装置

Publications (1)

Publication Number Publication Date
WO2020082396A1 true WO2020082396A1 (zh) 2020-04-30

Family

ID=70330641

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2018/112272 WO2020082396A1 (zh) 2018-10-27 2018-10-27 一种检测方法和检测装置

Country Status (2)

Country Link
CN (1) CN111356914B (zh)
WO (1) WO2020082396A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111739009A (zh) * 2020-06-23 2020-10-02 北京金山云网络技术有限公司 一种图像检测方法、装置、电子设备及存储介质

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114882028B (zh) * 2022-07-08 2022-10-21 深圳市瑞祥鑫五金制品有限公司 一种基于多摄像头的焊接端子检测方法、装置及系统
CN115953422B (zh) * 2022-12-27 2023-12-19 北京小米移动软件有限公司 边缘检测方法、装置及介质

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH01145556A (ja) * 1987-12-02 1989-06-07 Mitsubishi Heavy Ind Ltd グラフアイト製シールの非破壊品質保証法
CN101105442A (zh) * 2007-08-01 2008-01-16 北京交通大学 棉花中白色异物的多波长透反射检测方法及其装置
CN102495066A (zh) * 2011-12-05 2012-06-13 江南大学 高光谱透射图像采集系统及基于该系统的豆荚无损检测方法

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103247039B (zh) * 2013-05-09 2015-12-09 河海大学常州校区 一种基于复合视觉的高压线缆带电检测方法
US10115040B2 (en) * 2016-09-14 2018-10-30 Kla-Tencor Corporation Convolutional neural network-based mode selection and defect classification for image fusion

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH01145556A (ja) * 1987-12-02 1989-06-07 Mitsubishi Heavy Ind Ltd グラフアイト製シールの非破壊品質保証法
CN101105442A (zh) * 2007-08-01 2008-01-16 北京交通大学 棉花中白色异物的多波长透反射检测方法及其装置
CN102495066A (zh) * 2011-12-05 2012-06-13 江南大学 高光谱透射图像采集系统及基于该系统的豆荚无损检测方法

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111739009A (zh) * 2020-06-23 2020-10-02 北京金山云网络技术有限公司 一种图像检测方法、装置、电子设备及存储介质

Also Published As

Publication number Publication date
CN111356914B (zh) 2023-05-05
CN111356914A (zh) 2020-06-30

Similar Documents

Publication Publication Date Title
CN106097361B (zh) 一种缺陷区域检测方法及装置
TWI826586B (zh) 用於雜訊標籤的對抗訓練系統及方法
WO2020082396A1 (zh) 一种检测方法和检测装置
CN110390661A (zh) 检查管理系统、检查管理装置及检查管理方法
JP2020502490A (ja) 空気試料の自動分析を実行するためのシステム及び方法
CN110726724A (zh) 缺陷检测方法、系统和装置
CN107561737B (zh) 一种aoi系统ipc参数自动调整的方法、装置及工具
TWI614483B (zh) 光源亮度檢測方法及系統
JP2022507678A (ja) 自動目視検査プロセスにおけるセットアップ段階の最適化
CN115239684A (zh) 焊点检测方法、设备及存储介质
CN118089546B (zh) 一种图像分析监测药品稳定性的方法及系统
CN111344555B (zh) 评估泌乳哺乳动物的健康状态的方法和系统
CN105572133B (zh) 瑕疵检测方法及其装置
Jiang et al. Machine vision and background remover-based approach for PCB solder joints inspection
CN113096111A (zh) 物料检测方法、系统、计算机程序产品及可读存储介质
CN115165920A (zh) 一种三维缺陷检测方法及检测设备
US20210407070A1 (en) Methods and systems for non-destructive testing (ndt) with trained artificial intelligence based processing
JPH09264856A (ja) 物品外観検査装置
KR20200044172A (ko) 표면검사를 위한 딥러닝 학습방법
KR20220111214A (ko) 인공지능 기반 제품 결함 검사 방법, 장치 및 컴퓨터 프로그램
CN115147386A (zh) U型管的缺陷检测方法、装置及电子设备
KR20220101654A (ko) 자동화된 육안 검사 스테이션을 위한 오프라인 문제해결 및 개발
WO2020158630A1 (ja) 検出装置、学習器、コンピュータプログラム、検出方法及び学習器の生成方法
CN103901039A (zh) 检测系统
KR101748162B1 (ko) 능동형 비전 회로기판 검사장치 및 이를 이용한 회로기판 검사방법

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 18937789

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 08.09.2021)

122 Ep: pct application non-entry in european phase

Ref document number: 18937789

Country of ref document: EP

Kind code of ref document: A1