Automatic Defect Detection and Classification (ADC) Solution
Automatic Defect Detection and Classification (ADC) Solution
Akusense Sensor's Automatic Defect Classification (ADC) solution can comprehensively calculate and automatically classify defects generated during the production process, such as defect types, sizes, and locations;In the production and manufacturing process, traditional manual defect detection methods can no longer meet the detection needs, and intelligent detection technology is gradually playing an increasingly important role. Taking AOI (Automatic Optical Inspection) technology, which is widely used in the electronics industry, as an example, AOI first detects suspected defective products, and then manually identifies the type and location of defects. Although it has been integrated into automated production steps, it still has the inconvenience of time-consuming, labor-intensive, and costly detection processes.
Akusense Sensor's Automatic Defect Classification (ADC) solution can comprehensively calculate and automatically classify defects generated during the production process, such as defect types, sizes, and locations; Timely correction and improvement of some interfering factors, such as environmental interference, equipment failure, etc., to avoid the continued production of defective products. At the same time, it also provides guidance for subsequent rework, repair and other process operations, improves efficiency, reduces the defect rate of the entire system, timely reduces the workload of rework and repair, significantly improves defect identification rate, defect classification accuracy and detection efficiency.
This solution architecture has been applied to the defect detection of LCD panels in a leading semiconductor enterprise, achieving 99% defect recognition and automatic classification, and saving nearly 80% of labor costs.
Solution Value
increase rate
The defect recognition speed has been increased from 2-3 seconds/sheet for manual classification to 250ms/sheet, resulting in a 5% -6% increase in product yield and a less than 0.5% missed detection rate. This can save 80% of the labor cost for defect classification; Through defect type training and iteration, the highest defect recognition rate can reach over 99%.
Defect determination and classification
Through model training, the system has knowledge of defect discrimination, and ADC can automatically classify and discriminate defects, enabling real-time detection of various product defects with full coverage. In addition, the system simultaneously corrects erroneous results through sensing devices and provides real-time feedback to the ADC, allowing the model to continuously learn and improve defect discrimination performance.
Trained models can even detect tiny defects that are almost invisible to the naked eye. Engineers have skillfully used deep learning techniques to discover potential anomalies in microchips, transformers, displays, CPUs, and other manufacturing components - including many very small parts - and can also be applied to defect analysis of solar cell surfaces, PCB false point detection and analysis, and more.
+86 18188641602
Akusense Technology Co., Ltd.