Light source state detection system and light source state detection method

Abstract

The present application provides a light source state detection system and method, which comprises image acquisition, frame selection, light sign identification, marking, verification and light sign detection. The image acquisition acquires detection images. The frame selection selects frame selection areas of each target light source in each detection image. The light sign identification identifies the color and light state of each target light source. The marking marks the target light source serial number on each detection image. The verification generates an updated light source serial number when the light state of each target light source is off, and updates the target light source serial number to the updated light source serial number when the target light source serial number is different from the updated light source serial number. The light sign detection generates a detection result. In this way, the flicker frequency and flicker behavior of the target light source are detected and identified.

Description

Technical Field

[0001] This invention relates to a light source status detection system and a light source status detection method, and particularly to a light source status detection system and a light source status detection method on a circuit board. Background Technology

[0002] Circuit boards in electronic products have multiple light-emitting units. When the circuit on the circuit board is operating normally, the light-emitting units will emit light in a certain order, with a certain lighting duration and flashing frequency. Therefore, by detecting the light emission status of the light source on the circuit board, it is possible to determine whether the circuit is operating normally. However, currently, the light source status of the circuit board is detected manually by operators. When a circuit malfunction is found in subsequent processes, it is difficult to trace back to the light emission status of the light-emitting units, thus making it impossible to quickly find circuit errors based on the light emission status of individual light-emitting units.

[0003] In view of this, developing a light source state detection system and method that can simultaneously identify and record the luminous states of multiple target light sources has become a worthwhile research and development goal for relevant industries. Summary of the Invention

[0004] Therefore, the purpose of this invention is to provide a light source state detection system and method, which acquires images of multiple target light sources on a circuit board using multiple image acquisition devices and identifies their luminous states. Thus, the light source state detection system and method of this invention can quickly identify and record the light source states on a circuit board.

[0005] According to one embodiment of the system configuration of the present invention, a light source state detection system is provided, comprising a test frame, multiple image acquisition devices, and a processor. The test frame houses a circuit board. The circuit board includes multiple target light sources. The image acquisition devices are disposed on the test frame and correspond to the target light sources respectively. These image acquisition devices are used to acquire multiple detection images containing the target light sources. The processor is signal-connected to the image acquisition devices and receives the detection images. The processor is configured to perform operations including the following steps: a frame selection step, a light number recognition step, a marking step, a verification step, and a light number detection step. The frame selection step includes selecting a frame region for each target light source in each detection image. The light number recognition step includes identifying a color and a emission state of each target light source located in each detection image based on a deep learning algorithm. The marking step includes identifying the detection images based on another deep learning algorithm and marking each detection image with a target light source serial number. Each target light source corresponds to a target light source serial number. The verification step includes verifying the target light source number based on the selected areas in each detection image, and generating an updated light source number when the emission state of each target light source is off. If the target light source number and the updated light source number are different, the target light source number is updated to the updated light source number. The verification step also includes a loading step, an estimation step, a confirmation step, a judgment step, and a compensation step. The loading step includes loading multiple selected areas of multiple target light sources. The estimation step includes estimating the area of ​​a light source and the centroid coordinates of a light source in one of the multiple detection images. The confirmation step includes confirming whether the emission state of this light source in the multiple detection images is emitting light. The judgment step includes determining whether this light source in the multiple detection images has a corresponding target light source number when the emission state of this light source in the multiple detection images is off, and marking the target light source number when this light source in the multiple detection images does not have a corresponding target light source number. The compensation step includes re-identifying each detection image and generating an updated light source number when the emission state of this light source in the multiple detection images is off and this light source in the multiple detection images has a corresponding target light source number. The light indicator detection step includes generating a detection result based on each detection image corresponding to the target light source number. The detection result includes a flicker frequency and a flicker behavior for each target light source.

[0006] According to one embodiment of the method of the present invention, a light source state detection method is provided, comprising a setting step, an image acquisition step, a bounding selection step, a light number recognition step, a marking step, a verification step, and a light number detection step. The setting step includes setting a circuit board on a test fixture. The circuit board includes multiple target light sources. The image acquisition step includes driving multiple image acquisition devices to acquire multiple detection images including the target light sources. The bounding selection step includes driving a processor to bound each target light source within a selected area in each detection image. The light number recognition step includes driving the processor to identify a color and a emission state of each target light source located in each detection image according to a deep learning algorithm. The marking step includes driving the processor to identify the detection images according to another deep learning algorithm and mark each detection image with a target light source serial number. Each target light source corresponds to a target light source serial number. The verification step includes the driver processor verifying the target light source number based on the selected regions in each detection image, and generating an updated light source number when the emission state of each target light source is off. If the target light source number and the updated light source number are different, the target light source number is updated to the updated light source number. The verification step also includes a loading step, an estimation step, a confirmation step, a judgment step, and a compensation step. The loading step includes the driver processor loading multiple selected regions of multiple target light sources. The estimation step includes the driver processor estimating the area of ​​a light source and the centroid coordinates of a light source in one of the multiple detection images. The confirmation step includes the driver processor confirming whether the emission state of this light source in the multiple detection images is emitting light. The judgment step includes the driver processor determining whether this light source in the multiple detection images has a corresponding target light source number when the emission state of this light source in the multiple detection images is off, and marking the target light source number. The compensation step includes re-identifying each detection image and generating an updated light source number when the emission state of this light source in the multiple detection images is off and this light source in the multiple detection images has a corresponding target light source number. The light indicator detection step involves the driver processor generating a detection result based on each detection image corresponding to the target light source number. The detection result includes a flicker frequency and a flicker behavior for each target light source. Attached Figure Description

[0007] Figure 1 This is a schematic diagram illustrating a light source state detection system according to a first embodiment of the present invention;

[0008] Figure 2 This is a flowchart illustrating a light source state detection method according to a second embodiment of the present invention;

[0009] Figure 3 It is shown according to Figure 2 A flowchart of the verification steps for the light source state detection method;

[0010] Figure 4 It is shown according to Figure 2A flowchart of the compensation step in the verification process of the light source state detection method; and

[0011] Figure 5 It is shown according to Figure 2 The flowchart shows the lamp number detection steps of the light source status detection method.

[0012] Explanation of reference numerals in the attached figures:

[0013] 10: Circuit board

[0014] 100: Light Source Status Detection System

[0015] 110: Test fixture

[0016] 111: Positioning hole

[0017] 120: Image Acquisition Device

[0018] 130: Processor

[0019] S100: Light Source Status Detection Method

[0020] S01: Setup Steps

[0021] S02: Image Acquisition Steps

[0022] S03: Selection Steps

[0023] S04: Light Signal Recognition Steps

[0024] S05: Marking Steps

[0025] S06: Verification Steps

[0026] S061: Loading Steps

[0027] S062: Estimation Steps

[0028] S063: Confirmation Steps

[0029] S064: Judgment Step

[0030] S065, S066, S071, S072, S073, S074, S075: Steps

[0031] S067, S068, S069: Sub-steps

[0032] S07: Compensation Steps

[0033] S08: Light Signal Detection Procedure

[0034] S081: Statistical Steps

[0035] S082: Calculation Steps

[0036] CL: Color

[0037] Imt: Detection Image

[0038] L1, L2, L3: Target light source

[0039] R1: Detection Results

[0040] RG1: Selected area

[0041] SE: Target Light Source Number

[0042] ST: Illumination state Detailed Implementation

[0043] Please see Figure 1 , Figure 1 This is a schematic diagram illustrating a light source state detection system 100 according to a first embodiment of the present invention. The light source state detection system 100 includes a test fixture 110, a plurality of image acquisition devices 120, and a processor 130. The test fixture 110 houses a circuit board 10. The circuit board 10 includes a plurality of target light sources L1, L2, and L3. The image acquisition devices 120 are disposed on the test fixture 110 and correspond to the target light sources L1, L2, and L3 respectively. These image acquisition devices 120 are used to acquire a plurality of detection images Imt (see [reference needed]) containing the target light sources L1, L2, and L3. Figure 2 The processor 130 is signal-connected to the image acquisition device 120 and receives these detected images Imt. The processor 130 is configured to implement the light source state detection method S100 (see...). Figure 2 ).

[0044] Specifically, the test fixture 110 may include multiple positioning holes 111 located at different positions, which allow circuit boards 10 to be inserted. The image acquisition device 120 moves on the test fixture 110 according to the positions of the target light sources L1, L2, and L3. The processor 130 may be a microprocessor unit, a central processing unit, or other electronic processing unit, but the present invention is not limited thereto. Thus, the light source state detection system 100 of the present invention is equipped with multiple image acquisition devices 120, which can be used to detect multiple target light sources L1, L2, and L3 on electronic devices with complex structures or composed of multiple circuit boards 10 connected in different directions and angles.

[0045] Please see Figure 1 and Figure 2 ,in Figure 2This is a flowchart illustrating a light source state detection method S100 according to a second embodiment of the present invention. The light source state detection method S100 includes a setting step S01, an image acquisition step S02, a bounding selection step S03, a lamp number recognition step S04, a marking step S05, a verification step S06, and a lamp number detection step S08. The setting step S01 includes setting a circuit board 10 on a test fixture 110. The circuit board 10 includes target light sources L1, L2, and L3. The image acquisition step S02 includes driving an image acquisition device 120 to acquire multiple detection images Imt containing the target light sources L1, L2, and L3. The bounding selection step S03 includes driving a processor 130 to bound a region RG1 for each target light source L1, L2, and L3 within each detection image Imt. The lamp number recognition step S04 includes driving the processor 130 to identify a color CL and a light emission state ST for each target light source L1, L2, and L3 located in each detection image Imt based on a deep learning algorithm. The marking step S05 includes the drive processor 130 identifying the detection images Imt according to another deep learning algorithm and marking each detection image Imt with a target light source number SE. Each target light source L1, L2, L3 corresponds to the target light source number SE. The verification step S06 includes the drive processor 130 verifying the target light source number SE according to the selected area RG1 in each detection image Imt, and generating an updated light source number when the light emission state ST of each target light source L1, L2, L3 is off. When the target light source number SE is different from the updated light source number, the target light source number SE is updated to the updated light source number. The lamp number detection step S08 includes the drive processor 130 generating a detection result R1 according to each detection image Imt corresponding to the target light source number SE. The detection result R1 includes a flicker frequency and a flicker behavior of each target light source L1, L2, L3. In this way, the light source state detection method S100 of the present invention can quickly identify and record the light source state on the circuit board 10.

[0046] The light source state detection system 100 may further include at least one light guide element (not shown). The light guide element is connected to at least one of the target light sources L1, L2, and L3. The light source state detection method S100 may further include a light guiding step (not shown). The light guiding step connects the light guide element to at least one of the target light sources L1, L2, and L3, and extends a light emitted by at least one of the target light sources L1, L2, and L3. Therefore, when at least one of the target light sources L1, L2, and L3 is blocked by circuits or electronic components on the circuit board 10, the light guide element extends and guides the target light sources L1, L2, and L3 to a position where the image acquisition device 120 can smoothly acquire an image, for subsequent detection and identification.

[0047] In the image acquisition step S02, the image acquisition device 120 is positioned to acquire images containing the target light sources L1, L2, and L3.

[0048] In the selection step S03, the region RG1 containing one of the target light sources L1, L2, and L3 in the detection image Imt is selected.

[0049] In the light identification step S04, the processor 130 identifies the color CL and emission state ST (i.e., emitting or extinguishing) of the target light sources L1, L2, and L3 in all detected images Imt using a deep learning algorithm. The deep learning algorithm can be a feature extractor, a classifier, or a convolutional layer, but this invention is not limited thereto.

[0050] In the labeling step S05, the processor 130 uses another deep learning algorithm to determine the target light sources L1, L2, and L3 (e.g., target light source L1) corresponding to the detected image Imt, which has already identified the colors CL and emission states ST of the target light sources L1, L2, and L3, and labels the corresponding target light source sequence number SE in the current detected image Imt. The other deep learning algorithm can be a machine learning algorithm, an image processing algorithm, a model training algorithm, a classifier, or a Kalman filter, but this invention is not limited thereto.

[0051] Specifically, although the marking step S05 uses a deep learning algorithm to determine the target light source number SE corresponding to one of the target light sources L1, L2, and L3 in the detection image Imt, when the emission state ST of the target light sources L1, L2, and L3 is off, it may result in the inability to identify the target light sources L1, L2, and L3 corresponding to the current detection image Imt, or an incorrect judgment. Therefore, the light source state detection method S100 of the present invention further verifies the previously marked target light source number SE through a verification step S06 to ensure that the marking result is correct. The operation details of the verification step S06 will be described below through a more detailed embodiment.

[0052] Please refer to the following: Figures 1 to 3 ,in Figure 3 It is shown according to Figure 2 The flowchart illustrates the verification step S06 of the light source state detection method S100. Verification step S06 may further include a loading step S061, an estimation step S062, a confirmation step S063, and a judgment step S064. Loading step S061 includes the driver processor 130 loading multiple bounding boxes RG1 of the target light sources L1, L2, and L3. Specifically, in loading step S061, the bounding boxes RG1 corresponding to all target light sources L1, L2, and L3 selected in bounding step S03 are loaded into the detection image Imt.

[0053] The estimation step S062 includes driving the processor 130 to estimate the area of ​​a light source and the centroid coordinates of a light source in one of the detection images Imt. In other words, in the estimation step S062, the processor 130 calculates the area of ​​a light source and the centroid coordinates of a light source based on one of the target light sources L1, L2, and L3 in each detection image Imt.

[0054] The confirmation step S063 includes the drive processor 130 confirming whether the light emission state ST of each of the detection images Imt is emitting light. In the confirmation step S063, the processor 130 confirms whether the light emission state ST of each detection image Imt is emitting light or extinguished (i.e., not emitting light). If it is emitting light, step S065 is executed; if it is not emitting light (i.e., extinguished), the judgment step S064 is executed.

[0055] In step S065, when the emission state ST corresponding to the current detection image Imt is luminous, the light source area and the light source centroid coordinates estimated in step S062 are recorded, and the target light source number SE is marked on the current detection image Imt. In step S066, the target light source number SE and the emission states ST of target light sources L1, L2, and L3 are stored in the database.

[0056] The determination step S064 includes the driver processor 130 determining whether the detection images Imt have a corresponding target light source number SE when the light emission state ST of these detection images Imt is off, and marking the target light source number SE when the detection images Imt do not have a corresponding target light source number SE. Specifically, the determination step S064 includes sub-steps S067, S068, and S069. Sub-step S067 determines whether the detection images Imt with the light emission state ST off have been marked with a target light source number SE in the previous marking step S05. If not, sub-step S068 is executed; if so, compensation step S07 is executed.

[0057] Sub-step S068 calculates the light source area and the centroid coordinates of the light source in the detection image Imt, where the target light source number SE has not yet been marked. Sub-step S069 marks the target light source number SE of the aforementioned detection image Imt based on the aforementioned light source area and the centroid coordinates of the light source.

[0058] Please refer to the following: Figures 1 to 4 ,in Figure 4 It is shown according to Figure 2The flowchart below shows the compensation step S07 in the verification step S06 of the light source state detection method S100. The compensation step S07 includes re-identifying each detection image Imt when the light emission state ST of these detection images Imt is off, and when each of these detection images Imt has a corresponding target light source number SE, and generating an updated light source number. Specifically, the compensation step S07 includes steps S071, S072, S073, S074, and S075. Step S071 loads the aforementioned detection image Imt marked with the target light source number SE.

[0059] Step S072 performs the same operation as in step S05 on the current detection image Imt, identifying the target light source L1 corresponding to the current detection image Imt and generating an updated light source sequence number. Step S073 determines whether the currently generated updated light source sequence number is the same as the target light source sequence number SE generated in the previous marking step S05. If not, proceed to step S074; if yes, proceed to step S075. Step S074 updates the target light source sequence number SE corresponding to the current detection image Imt to the updated light source sequence number. Step S075 stores either the target light source sequence number SE or the updated light source sequence number into the database based on the determination result of step S073.

[0060] Please see Figures 1 to 2 and Figure 5 ,in Figure 5 It is shown according to Figure 2 The flowchart below shows the lamp number detection step S08 of the light source state detection method S100. The lamp number detection step S08 may further include a statistical step S081 and a calculation step S082. The statistical step S081 includes the drive processor 130 statistically analyzing the emission state ST of each detection image Imt corresponding to the target light source sequence number SE. The calculation step S082 includes the drive processor 130 calculating the flicker frequency and flicker behavior of the target light sources L1, L2, and L3 based on these emission states ST.

[0061] Specifically, in the statistical step S081, the emission state ST of all detected images Imt with the same target light source number SE is recorded sequentially. For example, the emission states ST of target light sources L1, L2, and L3 corresponding to target light source number SE are shown in Table 1.

[0062] Table 1

[0063]

[0064] In calculation step S082, the processor 130 calculates the flicker frequency of each target light source L1, L2, and L3 based on the emission state ST of each target light source L1, L2, and L3. Furthermore, calculation step S082 can further remove emission states ST whose flicker frequencies differ significantly from the standard deviation, and calculate the flicker frequencies of the target light sources L1, L2, and L3 for emission states ST within the standard deviation. Therefore, the light source state detection method S100 of the present invention can automatically filter out the detection images Imt of the target light sources L1, L2, and L3 in their initial state, and only detect the light source states of the target light sources L1, L2, and L3 after the circuit board 10 has entered normal operation.

[0065] In other embodiments of the present invention, the light source state detection method may further include a recording step. The recording step includes driving the processor to store the color, flicker frequency, and flicker behavior of each target light source into a database. Therefore, the light source state detection method of the present invention can retrospectively analyze the detection results of the light source state when a fault is detected in subsequent circuit board processes, thus accelerating the error finding process.

[0066] Although the present invention has been disclosed above with reference to embodiments, it is not intended to limit the present invention. Any person skilled in the art can make various changes and modifications without departing from the concept and scope of the present invention. Therefore, the scope of protection of the present invention shall be determined by the claims.

Claims

1. A light source status detection system, characterized in that, Include: A test fixture for mounting a circuit board, wherein the circuit board contains multiple target light sources; Multiple image acquisition devices are disposed on the test frame and respectively correspond to the multiple target light sources. These multiple image acquisition devices are used to acquire multiple detection images containing the multiple target light sources; and A processor, signal-connected to the plurality of image acquisition devices and receiving the plurality of detected images, is configured to perform operations including the following steps: A selection step includes selecting a selection area for each target light source in each detection image; A light identification step includes identifying the color and emission state of each target light source located in each detection image based on a deep learning algorithm; A labeling step includes identifying the plurality of detected images based on another deep learning algorithm and labeling each of the detected images with a target light source number, wherein each target light source corresponds to the target light source number; A verification step includes verifying the target light source number based on the selected area in each detection image, and generating an updated light source number when the emission state of each target light source is off. When the target light source number is different from the updated light source number, the target light source number is updated to the updated light source number. The verification step further includes: A loading step includes loading multiple selected regions of the multiple target light sources; An estimation step includes estimating the area of ​​a light source and the centroid coordinates of a light source among the plurality of detected images; The confirmation step includes confirming whether the luminescence state of the plurality of detected images is luminescent; A determination step includes determining whether the light emission state of the plurality of detection images is off when the light emission state of the plurality of detection images is off, and marking the target light source number when the plurality of detection images do not have the corresponding target light source number. A compensation step includes, when the light emission state of one of the plurality of detected images is off, and the one of the plurality of detected images has a corresponding target light source sequence number, re-identifying each of the detected images and generating the updated light source sequence number; and A light detection step includes generating a detection result based on each detection image corresponding to the target light source number, wherein the detection result includes a flashing frequency and a flashing behavior of each target light source.

2. The light source state detection system as described in claim 1, characterized in that, Also includes: At least one light guide element is connected to at least one of the plurality of target light sources and is used to extend a light emitted by the at least one of the plurality of target light sources.

3. The light source status detection system as described in claim 1, characterized in that, The light detection process also includes: A statistical step includes statistically analyzing the luminescence state of each detection image corresponding to the target light source sequence number; and One calculation step includes calculating the flicker frequency and flicker behavior of each target light source based on the light emission state.

4. The light source status detection system as described in claim 1, characterized in that, The processor is further configured to execute: A recording step includes storing the color, flicker frequency, and flicker behavior of each target light source into a database.

5. A method for detecting the state of a light source, characterized in that, Include: A setup step includes placing a circuit board on a test fixture, wherein the circuit board includes multiple target light sources; An image acquisition step includes driving multiple image acquisition devices to acquire multiple detection images containing the multiple target light sources; A selection step includes driving a processor to select a selection area for each target light source in each detection image; A light recognition step includes driving the processor to identify the color and emission state of each target light source located in each detection image based on a deep learning algorithm; A labeling step includes driving the processor to identify the plurality of detected images according to another deep learning algorithm and label each of the detected images with a target light source number, wherein each target light source corresponds to the target light source number; A verification step includes driving the processor to verify the target light source number based on the selected area in each of the detected images, and generating an updated light source number when the emission state of each target light source is off. When the target light source number is different from the updated light source number, the target light source number is updated to the updated light source number. The verification step further includes: A loading step includes driving the processor to load multiple selected regions of the multiple target light sources; An estimation step includes driving the processor to estimate the area of ​​a light source and the centroid coordinates of a light source in one of the plurality of detected images; A confirmation step includes driving the processor to confirm whether the luminescence state of the plurality of detected images is luminescent; A determination step includes driving the processor to determine whether the plurality of detected images have a corresponding target light source number when the light emission state of the image is off, and marking the target light source number. A compensation step includes, when the light emission state of one of the plurality of detected images is off, and the one of the plurality of detected images has a corresponding target light source sequence number, re-identifying each of the detected images and generating the updated light source sequence number; and A light detection step includes driving the processor to generate a detection result based on each detection image corresponding to the target light source number, wherein the detection result includes a flashing frequency and a flashing behavior of each target light source.

6. The light source state detection method as described in claim 5, characterized in that, Also includes: A light guiding step involves connecting at least one light guiding element to at least one of the plurality of target light sources, and using it to extend a light emitted by the at least one of the plurality of target light sources.

7. The light source state detection method as described in claim 5, characterized in that, The light detection process also includes: A statistical step includes driving the processor to statistically analyze the luminescence state of each detected image corresponding to the target light source sequence number; and One calculation step includes driving the processor to calculate the flicker frequency and flicker behavior of each target light source based on the light emission state.

8. The light source state detection method as described in claim 5, characterized in that, Also includes: A recording step includes driving the processor to store the color, flicker frequency, and flicker behavior of each target light source into a database.

Images