Currently, the application of AI technology in chip design is gradually expanding to the field of industrial visual inspection. DaoAI's AI AOI software system plays an important role in the detection of assembly missing and misassembly in the consumer goods/general industry.
User scenario: A leading consumer goods manufacturer. In the assembly process of its production line, the products involve the assembly of various components, and the inspection objects are all kinds of assembled finished products. During the production process, it is necessary to ensure that there are no missing or misassembled components in each product to guarantee the quality and performance of the products.
Pain points: In traditional assembly inspection, manual inspection is inefficient, with a missed detection rate as high as 4% and a false alarm rate of 30%. With the successful application of AI technology in fields such as chip design, the demand for intelligent inspection in industrial visual inspection is increasing. At the same time, the manufacturer needs to spend a lot of time reprogramming and debugging when changing product models. The model-change time is as long as 3 hours, which seriously affects production efficiency and makes it difficult to meet compliance requirements.
Technical principle
The AI AOI software system uses a visual foundation model for feature recognition. Its core lies in extracting and analyzing features in images through advanced algorithms. In the detection of assembly missing and misassembly, the system uses the APDT positive-sample/few-sample learning algorithm. Only 1-20 good product images are needed as samples, and the system can quickly learn the normal feature patterns of the products.
- For feature recognition, the system is based on a deep-learning network and can automatically identify features such as the shape, position, and color of different components. For example, by learning from good product images, the system remembers the standard shape and installation position of a certain screw. When the shape of the screw in the detected image is abnormal or the installation position is incorrect, the system can identify it in time.
- The semantic false-alarm filtering algorithm can effectively reduce false alarms. During the detection process, the system further judges the detection results based on semantic information. For instance, when a suspected defect area is detected, but the semantic information of this area matches the normal product features, the system filters out this false alarm.
- The 0-code automatic programming function enables the system to complete programming based on one good product in 5 minutes. This greatly shortens the programming time and improves the model-change efficiency. The principle is that the system automatically analyzes the features of the good product image and converts them into detection rules without the need for manual writing of complex codes.
DaoAI Solution and Product Introduction
DaoAI's AI AOI software system is the core product to solve the problems of assembly missing and misassembly. The system has the feature-recognition ability of the visual foundation model and can quickly and accurately identify the normal and abnormal features of products. Through APDT positive-sample/few-sample learning, only a small number of good-product samples are needed to complete model training, which greatly reduces the workload of sample collection and annotation.
The AI AOI software system brings a new solution to industrial quality inspection with its efficient programming and learning capabilities.
The semantic false-alarm filtering function of the system effectively reduces the false-alarm rate and improves the accuracy of detection. At the same time, it supports 100% local private deployment of SDK/API/Docker, and the data does not leave the factory, ensuring the security and privacy of data. In practical applications, it can be used in conjunction with the DaoAI 2D/3D AI AOI equipment, using its self-developed 3D camera and three-dimensional morphology reconstruction technology to conduct more comprehensive inspections on products.
Quantifiable results: After using the AI AOI software system, the manufacturer's detection rate has increased to 99.2%, the missed-detection rate has decreased to <0.8%, the false-alarm rate has decreased by -70%, and the model-change time has been shortened from 3 hours to 5 minutes, greatly improving production efficiency and product quality.
FAQ
How many good-product samples are needed for the training of the AI AOI software system?
The system uses the APDT positive-sample/few-sample learning algorithm. Only 1-20 good-product samples are needed to complete the training, which greatly reduces the workload of sample collection and annotation.
What is the model-change time of the system?
The system has a 0-code automatic programming function. Based on one good product, programming can be completed in 5 minutes, which significantly shortens the model-change time and improves production efficiency.
How does the system reduce the false-alarm rate?
The system uses a semantic false-alarm filtering algorithm. It further judges the detection results based on semantic information and filters out suspected defects that match the normal product features, effectively reducing the false-alarm rate.