Defective Judgment Learning System

basic information

Sensors, collection and accumulation of measurement data, capable of storing time-series data. Supports management of image files, CSV files, etc. AI learns from the collected measurement data and good/bad judgment result data. Using the trained AI, good/bad judgment of measurement data is determined in real-time. Good/bad judgment results are also saved to the database, and judgment history is displayed on the screen. We support customized development tailored to the on-site environment, such as measurement data and judgment methods. It is also possible to add information such as defect information and countermeasure information during failures, as well as search for similar cases, which can be useful for analyzing trends and considering countermeasures when defects occur.

Price range

Delivery Time

Applications/Examples of results

It flexibly accumulates and manages data according to various formats of sensors and measurement data. It can handle large amounts of time-series data as well. Files in data formats that are difficult to directly input into images or databases are saved as files and managed within the database. It can also be utilized as a storage and management system for diverse measurement data and files. Data from sensors of processing machines and measurement devices, along with the judgment results of good and defective products, are registered as training data for AI and used for learning. Pre-trained AI can be saved for each product type or project. During operation, data from the sensors of processing machines and measurement devices is received, and the pre-trained AI performs real-time judgment of product quality, returning the results. The judgment status and history can be checked on the web interface. Information on defects and countermeasures that occur during issues can be registered in the database in conjunction with the defective judgment data. It is also possible to register documents such as defect reports. Keywords can be automatically extracted and registered from documents, allowing for flexible searching. In the event of defects, it is possible to view past defect and countermeasure histories or search for similar cases in different products.