Core technology: One-click programming
Through cutting-edge deep learning algorithms, Maker-ray revolutionizes AOI programming with innovative one-click programming. This technological innovation addresses the challenges associated with traditional AOI programmings, such as time-consuming processes and complex operations. The optimization driven by AI simplifies the traditional automated visual inspection, making it accessible to all, irrespective of engineering expertise. Furthermore, AI-driven optimization reduces false alarms caused by manual programming and debugging errors while boosting programming efficiency by 50%. Therefore, AI-powered automated visual inspections are ideal vision systems for manufacturing.
Component inspection:
- Automation: search for nearly 100 components types (keep updating), identify polarity of capacitors, sockets, diodes and other components automatically.
- Excellent recognition ability: even the black polarity of black materials can be identified by multi-dimensional grabbing features;
- Strong anti-interference ability: clearly distinguishing component screen printing and board screen printing, reducing the interference of screen printing on programming；
- Operation simplified: The detection window (main body detection frame, sub-detection frame, solder detection frame) is automatically displayed.
- Intelligence: matching algorithm, adjust parameters, display pre-judgment results ntelligently(such as intelligent offset algorithm, OCR identification algorithm, missing parts etc)
Solder joint inspection:
AI intelligent algorithm searches the solder joints of whole board automatically, calculates position of components accurately, distinguishes solder joints needed be detected, and configures the algorithm parameters automatically. Easy operating, short programming time (15-20min).
Core technology: One-click programming
Through cutting-edge deep learning algorithms, Maker-ray revolutionizes AOI programming with innovative one-click programming. This technological innovation addresses the challenges associated with traditional AOI programmings, such as time-consuming processes and complex operations. The optimization driven by AI simplifies the traditional automated visual inspection, making it accessible to all, irrespective of engineering expertise. Furthermore, AI-driven optimization reduces false alarms caused by manual programming and debugging errors while boosting programming efficiency by 50%. Therefore, AI-powered automated visual inspections are ideal vision systems for manufacturing.
Component inspection:
- Automation: search for nearly 100 components types (keep updating), identify polarity of capacitors, sockets, diodes and other components automatically.
- Excellent recognition ability: even the black polarity of black materials can be identified by multi-dimensional grabbing features;
- Strong anti-interference ability: clearly distinguishing component screen printing and board screen printing, reducing the interference of screen printing on programming；
- Operation simplified: The detection window (main body detection frame, sub-detection frame, solder detection frame) is automatically displayed.
- Intelligence: matching algorithm, adjust parameters, display pre-judgment results ntelligently(such as intelligent offset algorithm, OCR identification algorithm, missing parts etc)
Solder joint inspection:
AI intelligent algorithm searches the solder joints of whole board automatically, calculates position of components accurately, distinguishes solder joints needed be detected, and configures the algorithm parameters automatically. Easy operating, short programming time (15-20min).
Core technology: AI training model
The AI training model keeps device updated and iterated rapidly. The more being used, the more powerful it is, because its database keep increasing by using more. Through image acquisition-standard-training, the accurate model can be obtained, which is helpful for the detection of incorrect position and uncommon defects. With simplified operation and strong adaptability, it can be used instantly after training. And it protects the privacy of factory process data.
Core technology: AI training model
The AI training model keeps device updated and iterated rapidly. The more being used, the more powerful it is, because its database keep increasing by using more. Through image acquisition-standard-training, the accurate model can be obtained, which is helpful for the detection of incorrect position and uncommon defects. With simplified operation and strong adaptability, it can be used instantly after training. And it protects the privacy of factory process data.
Core technology: OCR algorithm
In the SMT package inspection, the common way to control the wrong material and wrong part by detecting the silk screen characters.
The detection effect is easily affected by following aspects:
- Incoming itself affected by the process and differences existed in color and printing;
- Devices are polluted on the assembly line, resulting in dirty silk screens;
- During the process, some substitute materials for the same device, the font content is the same, but the shape and size of the font are different; or the substitute materials is with entirely different character content
- Screen printing recognition in dark scenes is close to character recognition problems (6-8-9), (2-5), (3-E), etc.
Based on deep learning algorithm, Maker-ray AOI adopts OCR recognition + intelligent positioning solution, which is an “end-to-end” algorithm. Image input, string of the recognition result output. With strong generalization, human intervention is non-required in process, which can eliminate dirt and silk screen interference automatically, recognize the real characteristics of silk screen characters intelligently. Thus, it is more efficient than the traditional OCR algorithm.
The biggest advantage of intelligent positioning method is that detection window adjustment is not needed. As long as the detection window is within a reasonable range, the characters can be framed and recognized correctly. OCR recognition effect is guaranteed. Programming efficiency is improved.
Core technology: OCR algorithm
In the SMT package inspection, the common way to control the wrong material and wrong part by detecting the silk screen characters.
The detection effect is easily affected by following aspects:
- Incoming itself affected by the process and differences existed in color and printing;
- Devices are polluted on the assembly line, resulting in dirty silk screens;
- During the process, some substitute materials for the same device, the font content is the same, but the shape and size of the font are different; or the substitute materials is with entirely different character content
- Screen printing recognition in dark scenes is close to character recognition problems (6-8-9), (2-5), (3-E), etc.
Based on deep learning algorithm, Maker-ray AOI adopts OCR recognition + intelligent positioning solution, which is an “end-to-end” algorithm. Image input, string of the recognition result output. With strong generalization, human intervention is non-required in process, which can eliminate dirt and silk screen interference automatically, recognize the real characteristics of silk screen characters intelligently. Thus, it is more efficient than the traditional OCR algorithm.
The biggest advantage of intelligent positioning method is that detection window adjustment is not needed. As long as the detection window is within a reasonable range, the characters can be framed and recognized correctly. OCR recognition effect is guaranteed. Programming efficiency is improved.
Core technology: 3D imaging & measurement technology
- Using multi-directional structured light to solve the problem of imaging deformation caused by shadow occlusion and ensure real imaging.
- The four-direction structured light is projected on the board, and the moiré light is reflected on the device. Three-dimensional information is obtained by three-dimensional imaging methods such as multi-frequency four-step phase shift algorithm and phase height mapping algorithm, and 3D images are produced by combining point cloud images.
- Intelligent depth map denoising deal with noise caused by ambient light and avoid noise from interfering with height measurement.
- Accuracy calibration byusing phase height mapping algorithm
Highly repeatable measurement accuracy
Cmk(capability machine index）stands for equipment detection repeatability.
AOI CMK of Maker-ray is greater than 20.
(Industry requirement is generally not less than 1.67. The larger the value, the higher the stability. )
Core technology: 3D imaging & measurement technology
- Using multi-directional structured light to solve the problem of imaging deformation caused by shadow occlusion and ensure real imaging.
- The four-direction structured light is projected on the board, and the moiré light is reflected on the device. Three-dimensional information is obtained by three-dimensional imaging methods such as multi-frequency four-step phase shift algorithm and phase height mapping algorithm, and 3D images are produced by combining point cloud images.
- Intelligent depth map denoising deal with noise caused by ambient light and avoid noise from interfering with height measurement.
- Accuracy calibration byusing phase height mapping algorithm
Highly repeatable measurement accuracy
Cmk(capability machine index）stands for equipment detection repeatability.
AOI CMK of Maker-ray is greater than 20.
(Industry requirement is generally not less than 1.67. The larger the value, the higher the stability. )
Get A Quote Today