Search

CN-121978129-A - Automatic detection system for lamp bead missing of LED display screen based on machine vision

CN121978129ACN 121978129 ACN121978129 ACN 121978129ACN-121978129-A

Abstract

The invention belongs to the technical field of LED display screens, and particularly relates to an automatic detection system for lamp bead missing of an LED display screen based on machine vision. This LED display screen lamp pearl lacks automated inspection system based on machine vision, through setting up drive part, this clamping part has realized the flexible self-adaptation fixed and the harmless centre gripping to the LED display screen, specifically, the crane makes contact gasbag contact screen body under drive arrangement drives, through filling clean gas after filtering, but a plurality of contact gasbag self-adaptation laminating different thickness or slightly have the display screen surface of deformation, evenly disperses pressure, has effectively avoided the screen body crush injury that traditional rigid fixture probably caused, scratch or the microcrack that stress concentration leads to.

Inventors

  • LI CHENG
  • NI JUN
  • LIU LINYU

Assignees

  • 安徽凝彩新型显示器件有限公司

Dates

Publication Date
20260505
Application Date
20260210

Claims (10)

  1. 1. The automatic detection system for the lamp bead loss of the LED display screen based on the machine vision comprises a conveying line (1) and is characterized in that a detection component is arranged on the outer surface of a frame of the conveying line (1), the detection component comprises a detection camera (22), and the detection camera (22) detects the lamp bead loss of the LED display screen; The LED display screen detection device comprises a conveying line (1), wherein a clamping part is fixedly arranged on the outer surface of a frame of the conveying line (1), the clamping part comprises a driving device and a fixing device, the driving device comprises a lifting frame (32), the lifting frame (32) lifts and lowers the fixing device, the fixing device comprises a contact air bag (54), and the contact air bag (54) is used for fixing an LED display screen to be detected; the system also comprises a detection system, wherein the detection system comprises a multi-physical-field sensing unit, an intelligent analysis and judgment unit, an actuator and feedback unit and a system monitoring and early warning unit.
  2. 2. The automatic detection system for the lamp bead loss of the LED display screen based on machine vision according to claim 1, wherein the detection component further comprises a detection outer frame (2), the detection outer frame (2) is fixedly installed on the upper surface of a frame of the conveying line (1), a moving component (21) is fixedly installed on the inner wall of the detection outer frame (2), the lower surface of a sliding block of the moving component (21) is fixedly installed on the outer surface of the detection camera (22), and an illumination component (23) is fixedly installed on the outer surface of the sliding block.
  3. 3. The automatic detection system for the lamp bead loss of the LED display screen based on machine vision of claim 2, wherein the driving device further comprises a lifting hydraulic cylinder (3), the lifting hydraulic cylinder (3) is fixedly arranged on the outer surface of the frame of the conveying line (1), a pushing frame (31) is fixedly arranged at one end of a piston rod of the lifting hydraulic cylinder (3), the outer surface of a limiting rod of the pushing frame (31) is slidably inserted into the inner wall of the lifting frame (32), a tension spring (33) is fixedly arranged on the outer surface of the limiting rod of the pushing frame (31), one end of the tension spring (33) is fixedly arranged on the outer surface of the lifting frame (32), a limiting plate (34) is fixedly arranged on the outer surface of the frame of the conveying line (1), and the lower surface of the limiting plate (34) is in contact with the upper surface of the lifting frame (32).
  4. 4. The automatic detection system for lamp bead loss of the LED display screen based on machine vision according to claim 3, wherein a limit guide rail (4) is fixedly arranged on the upper surface of the lifting frame (32), a movable frame (41) is slidably inserted on the upper surface of the limit guide rail (4), limit balls (42) are rotatably connected to the lower surface of the movable frame (41) through bearings, a guide groove plate (43) is fixedly arranged on the outer surface of the lifting frame (32), and the outer surface of the limit balls (42) is slidably connected with the inner wall of the guide groove plate (43).
  5. 5. The automatic detection system for the lamp bead loss of the LED display screen based on machine vision of claim 4, wherein the fixing device further comprises an air conveying air pump (5), the air conveying air pump (5) is fixedly installed on the outer surface of a frame of the conveying line (1), a filter (51) is fixedly installed on the outer surface of the frame of the conveying line (1), the air outlet end of the filter (51) is fixedly communicated with the air inlet end of the air conveying air pump (5), the air outlet end of the air conveying air pump (5) is fixedly communicated with a split flow pipeline (52) with a control valve, the air outlet air pump (53) is fixedly installed on the outer surface of the frame of the conveying line (1), and the outer surface of the split flow pipeline (52) is fixedly communicated with the air inlet end of the air outlet air pump (53) through a three-way valve.
  6. 6. The machine vision-based automatic detection system for lamp bead loss of an LED display screen, as set forth in claim 5, wherein the outer surfaces of the contact air bags (54) are fixedly installed on the outer surfaces of the movable frame (41) and the lifting frame (32), telescopic air bag groups (55) are fixedly installed on the inner walls of the contact air bags (54), the telescopic air bag groups (55) are fixedly communicated through connecting hoses, conveying hoses (56) are fixedly communicated on the outer surfaces of the telescopic air bag groups (55), one end of each conveying hose (56) is fixedly communicated with one end of each shunt pipeline (52), and two groups of conveying hoses (56) are respectively located inside the lifting frame (32) and the pushing frame (31).
  7. 7. The machine vision-based automatic detection system for lamp bead deficiency of an LED display screen of claim 6, wherein the multi-physical-field sensing unit comprises a multi-physical-field sensor array and a signal fusion and feature extraction module; The multi-physical field sensor array comprises a terahertz scanning imaging sensor, a laser induced ultrasonic sensor and a hyperspectral microscopic imaging sensor, and deep information of a screen is acquired from three physical fields of electromagnetic waves, mechanical vibration and material spectrum; the receiving and transmitting probe array of the terahertz scanning imaging sensor is arranged on the detection outer frame (2), and performs area array scanning on the LED display screen stagnated at the detection station to acquire chip structure information below the packaging layer; the pulsed laser emitter and the laser interferometer of the laser-induced ultrasonic sensor are integrated on the moving assembly (21), move along with the moving assembly and aim at each lamp bead unit, and excite and detect microscopic vibration spectrums of the lamp bead units; the hyperspectral microscopic imaging sensor is integrated at the position of the detection camera (22) and acquires microscopic reflection spectrum characteristics of each lamp bead area in a wide spectrum range; The signal fusion and feature extraction module is used for carrying out synchronization, filtering and noise reduction treatment on the original data acquired by the three sensors, extracting key features including terahertz reflection intensity images, ultrasonic feature spectrums and material reflection spectrums, and providing stable and multidimensional fusion data sources for subsequent judgment.
  8. 8. The machine vision-based automatic detection system for lamp bead deficiency of an LED display screen of claim 7, wherein the intelligent analysis and judgment unit comprises a core analysis module and a multi-mode defect judgment algorithm module; The core analysis module receives the fusion characteristic data from the multi-physical field sensing unit and executes analysis logic; The multi-modal defect judgment algorithm module is internally provided with a multi-source information fusion network and an expert rule base based on deep learning, and is used for calculating the physical existence confidence and internal health score of each lamp bead position according to the fused multi-physical field characteristic data, dynamically generating a defect type judgment result by comparing a preset defect characteristic mapping model, wherein the judgment result comprises complete missing, chip breakage, internal rosin joint and gold wire breakage, and realizing accurate identification and classification of defects; The multi-modal defect judgment algorithm module realizes accurate identification and classification of defects through the following mathematical process: The multi-source confidence fusion calculation comprises the steps of firstly, calculating independent confidence according to each sensor signal: , wherein, For the sensor type index to be used, Is a sensor The characteristic values of the original signals are collected and preprocessed, 、 、 Respectively terahertz reflection intensity, ultrasonic vibration amplitude and hyperspectral characteristic value, For corresponding sensor Is used for the decision threshold of (a), Is a sensor Is used for the sensitivity coefficient of the (c) for the optical sensor, Is a natural constant which is used for the production of the high-temperature-resistant ceramic material, Sensor-based Independent confidence of signal calculation; then, a sensor reliability factor is introduced , wherein, In order for the sensor to operate at a temperature, And For its optimal operating temperature point and stability parameters, Is a sensor Is a real-time reliability weighting factor for (1); Finally, calculate the integrated presence confidence of the lamp beads: , wherein, Is a sensor Dynamic weight coefficients in fusion satisfy The value of the sensor signal quality and the historical effect can be adaptively adjusted; defect pattern classification, construction of feature vectors ; Wherein, the Is the degree of difference of the signals, wherein, An integrated feature vector for defect classification, Is a sensor And (3) with The degree of signal difference between them, In order to represent the transpose of the vector, For the independent confidence of the terahertz sensor, For the independent confidence of the laser ultrasonic sensor, For the independent confidence of the hyperspectral sensor, For the degree of difference between terahertz and laser ultrasound signals, For the degree of difference between the laser ultrasound and the hyperspectral signal, For the degree of difference between hyperspectral and terahertz signals, Is a sensor Is used for the normalized feature value of (a); Will be Inputting a multi-classifier based on a Softmax function, and calculating the classification of the defects Probability of (2): , wherein, For a given feature vector When the detected lamp beads belong to the defect category Is a function of the conditional probability of (1), For the category index to be used, To express the first A class of defects is defined as a group of defects, The index is summed up for the defect class, And For and category of The corresponding weight vector and bias term(s), The total number of defect categories preset for the system, Is that In the form of a row vector of (c), Is that In the form of a row vector of (c), Is the first The weight vector of the class defect, Is the first Class bias items; Parameter self-adaptive adjustment, namely, according to feedback of historical judgment results, minimizing a loss function Dynamically optimizing the decision threshold using gradient descent Weight coefficient , Is the first Individual history samples for categories Is a real tag of the (c) in the (c), For a batch of historical sample numbers for parameter optimization, Index samples in the historical sample lot.
  9. 9. The machine vision-based automatic detection system for lamp bead loss of an LED display screen of claim 8, wherein the actuator and feedback unit comprises a report generation and marking triggering module and a system self-checking and status feedback module; The report generation and marking triggering module is used for receiving the defect coordinate and type information from the intelligent analysis and judgment unit, automatically generating a structured detection report, driving external marking equipment to mark the position of the screen defect, or sending a sorting instruction to the production line control system; The system self-checking and state feedback module is used for automatically calibrating the reference parameters of the multi-physical-field sensor before the detection starts or in an idle period, and feeding back the working states of the sensors, the moving assembly (21) and the lighting assembly (23) to a monitoring center in real time.
  10. 10. The machine vision-based automatic detection system for lamp bead missing of an LED display screen of claim 9, wherein the system monitoring and early warning unit comprises a comprehensive fault diagnosis module and a grading early warning and indication module; The comprehensive fault diagnosis module is used for monitoring system level or process level faults such as abnormal sensor signals, data flow interruption, low analysis algorithm confidence, overtime response of an actuator and the like, and generating a diagnosis log containing fault level and positioning information; the comprehensive fault diagnosis module calculates and judges the uncertainty measure To monitor the consistency between the sensors when Triggering a self-test when a preset threshold is exceeded, wherein, Is the average confidence; When the system fault is diagnosed, or the screen defect rate is detected to exceed a process threshold value, or the high risk defect type is determined, the grading early warning and indication module triggers grading warning through a remarkable color block and sound of a local human-machine interface (HMI) and a connected central control console of a production line, and pauses the conveying line (1) to be processed.

Description

Automatic detection system for lamp bead missing of LED display screen based on machine vision Technical Field The invention relates to the technical field of LED display screens, in particular to an automatic detection system for lamp bead missing of an LED display screen based on machine vision. Background The LED display screen is widely applied to a plurality of fields such as advertisement media, information release, stage performance, command scheduling and the like, and the display quality and reliability of the LED display screen are directly related to the use effect and the user experience. The LED display screen is formed by arranging tens of thousands or even millions of miniature LED lamp beads (pixel points) in a matrix mode, and defects such as complete deletion, physical damage, internal cold joint, gold wire fracture and the like of individual lamp beads can occur in the production, transportation or long-term use process, so that the display screen is subjected to dark spots, color cast or influence on the uniformity of the whole picture, and local or large-area failure is even caused in severe cases. Therefore, the LED display screen is subjected to rapid and accurate lamp bead defect detection before delivery or in a maintenance link, and the method is a key procedure for guaranteeing the product quality. The traditional defect detection mainly relies on manual visual inspection, and the method has low efficiency, high labor intensity and is easily influenced by subjective factors of personnel, and potential welding or structural defects inside the lamp beads are difficult to find, so that the requirements of modern mass production on efficiency and consistency cannot be met. In order to improve the detection automation level, automatic detection technology based on machine vision is gradually applied. The common scheme is that a display screen is sent to a detection station through a conveying line, a screen is electrified to be lightened (such as a full white field or a specific test pattern is displayed), an industrial camera is utilized to collect a display picture, and a non-bright or abnormal lamp bead is identified through an image processing algorithm (such as comparison with a standard template and brightness/chromaticity analysis). The method has the obvious limitations that firstly, the display screen is electrified, the complexity of butt joint of an electric interface and the risk of poor contact are increased, the detection cannot be carried out under the condition that the screen cannot be lightened (such as a power module fault), secondly, the detection result is extremely easy to be interfered by factors such as ambient light, reflection of the surface of the screen, different display contents and the like, the false detection rate is higher, thirdly, the method can only judge whether the lamp bead is bright or not bright, and is difficult to make effective judgment under the conditions of 'sub-health' state that the electric connection is unstable but not completely invalid due to poor internal welding, chip microcrack, gold wire breakage and the like and the condition that the lamp bead is physically present but the internal structure is damaged. In recent years, some improved techniques, such as detection of heat generation of the lamp beads after power-on by infrared thermal imaging, have been developed. The method can provide certain internal information, still needs to be energized and excited, and the thermal response is greatly influenced by the heat dissipation condition and is insensitive to small thermal differences. In addition, there are studies on how to realize synchronous and reliable identification of multiple types of defects in a rapid, high-precision and on-line detection scene of the LED display screen lamp beads by adopting a non-contact flaw detection technology (such as ultrasonic) for electronic element detection, and a mature and effective systematic solution is still lacking. Disclosure of Invention Based on the technical problems that an existing automatic detection system for the lamp bead loss of an LED display screen is low in efficiency, high in labor intensity and easy to influence by subjective factors of personnel, potential welding or structural defects inside the lamp bead are difficult to find, the lamp bead is physically present but the internal structure is damaged, and effective discrimination is difficult to make, the invention provides an automatic detection system for the lamp bead loss of the LED display screen based on machine vision. The invention provides an automatic detection system for the missing of LED display screen beads based on machine vision, which comprises a conveying line, wherein a detection part is arranged on the outer surface of a frame of the conveying line, the detection part comprises a detection camera, and the detection camera detects the missing of the LED display screen beads; The clamping component