Real-time process defect detection automation system and method using machine learning model
An artificial intelligence-based process defect detection system may include: a photographing module that collects image data by capturing a process that progresses on an object; a machine learning model that generates work data that is a result of recognizing and reading the object based on the image data; and a detection module that receives instruction data recorded regarding a process for an object optimized for product production, detects a defect or a non-defect by comparing the work data with the instruction data, and generates defect information when the process is defective.
1 . An artificial intelligence-based process defect detection system comprising:
a photographing module that collects image data by capturing a process that progresses on an object;
a machine learning model that generates work data that is a result of recognizing and reading the object based on the image data;
a detection module that receives instruction data recorded regarding a process for an object optimized for product production, detects a defect or a non-defect by comparing the work data with the instruction data, and generates defect information when the process is defective;
a reading module evaluates each unit process according to an evaluation standard, and accumulates and evaluates the unit processes in order, generating execution data that is collected and transmitted to a determination module;
the determination module determines the optimal execution from among multiple executions performed by changing the order of the unit processes based on the execution data of the reading module;
the optimal execution is determined by the determination module to be an execution in which the total required time or defect rate of the unit process recorded in the execution data is minimum;
where in the process that progresses on the object includes unit processes having a series of sequences,
a work module in which a work area where the object is processed is located is further included,
the machine learning model generates work data including location information of the work module by detecting the work area of the work module,
the machine learning model recognizes that the object changes as each of the unit processes progresses, and generates work data including sequence information obtained by reading the sequence of the process,
the machine learning model recognizes an outer appearance of the object that changes as the process progresses, and generates work data including state information obtained by reading a work state of the process, and
the detection module generates defect information including the work state, the sequence of the progress, and the work area by comparing the work data with the instruction data;
where in the machine learning model measures similarity to the image data based on normal image data collected in advance for the object, and generates work data including state information obtained by reading the work state of the unit process on the image data using a predefined similarity criterion,
a searching module that searches for an object related to the object when there is no normal image data collected in advance for the object is included,
the similarity to the image data is measured using the normal image data collected in advance for an object related to the object searched for by the searching module, and
the detection module detects state information of the work data and state information of the instruction data as being normal when the state information of the work data and the state information of the instruction data are determined to be similar to each other using a predefined error criterion, and generates defect information by detecting the state information of the work data and the state information of the instruction data as a state defect when the state information of the work data and the state information of the instruction data are different from each other.
2 . An artificial intelligence-based process defect detection method comprising:
a step in which a capturing module inputs, to a machine learning model, image data generated by capturing a process that progresses on an object;
a step in which the machine learning model recognizes the object using the image data, generates work data for the progressed process, which is a result of reading the image data, and transmits the work data to the detection module; and
a step where the detection module detects the presence or absence of defect by comparing the work data with instruction data obtained by recording an optimization process for a product produced by processing the object;
a reading module evaluates each unit process according to an evaluation standard, and accumulates and evaluates the unit processes in order, generating execution data that is collected and transmitted to a determination module;
the determination module determines the optimal execution from among multiple executions performed by changing the order of the unit processes based on the execution data of the reading module;
the optimal execution is determined by the determination module to be an execution in which the total required time or defect rate of the unit process recorded in the execution data is minimum;
where in the process that progresses on the object includes unit processes having a series of sequences,
a work module in which a work area where the object is processed is located is further included,
the machine learning model generates work data including location information of the work module by detecting the work area of the work module,
the machine learning model recognizes that the object changes as each of the unit processes progresses, and generates work data including sequence information obtained by reading the sequence of the process,
the machine learning model recognizes an outer appearance of the object that changes as the process progresses, and generates work data including state information obtained by reading a work state of the process, and
the detection module generates defect information including the work state, the sequence of the progress, and the work area by comparing the work data with the instruction data;
wherein the machine learning model measures similarity to the image data based on normal image data collected in advance for the object, and generates work data including state information obtained by reading the work state of the unit process on the image data using a predefined similarity criterion,
a searching module that searches for an object related to the object when there is no normal image data collected in advance for the object is included,
the similarity to the image data is measured using the normal image data collected in advance for an object related to the object searched for by the searching module, and
the detection module detects state information of the work data and state information of the instruction data as being normal when the state information of the work data and the state information of the instruction data are determined to be similar to each other using a predefined error criterion, and generates defect information by detecting the state information of the work data and the state information of the instruction data as a state defect when the state information of the work data and the state information of the instruction data are different from each other.