A method of inspection of a real physical object potentially affected by real anomalies, said method comprising the following steps:
- a) defining an interpretation model (IM) for evaluating the representativeness of said real physical object or of said real anomalies in the real images or in the virtual images;
- b) obtaining a real acquisition system comprising the following components:
- one or more illuminators for generating and directing radiation towards a real physical object potentially affected by anomalies or defects;
- one or more cameras for acquiring real images of said object by means of said radiation;
- one or more background elements interacting with said radiation and conditioning the acquisition process of said real images;
- a processing unit operating a computer program (SW) for processing the acquired images,
- c) obtaining a simulation engine capable of generating a digital clone (DT) of said real acquisition system in a simulated environment (SE) dual of the real environment (RE), said digital clone (DT) comprising input models which include:
- an object digital model virtual clone of said ideal physical object or of said real physical object;
- a defect digital model virtual clone of said real anomalies or real defects;
- one or more simulated cameras virtual clones of said one or more cameras;
- one or more simulated illuminators virtual clones of said one or more illuminators;
- a digital model of background elements virtual clone of said background elements present in the real environment (RE),
- said models being defined by an architecture and by parameters describing their configuration and behavior in the virtual environment (SE);
- d) applying the digital model of the defect to said digital model of object, in order to obtain a model of the object which contains information on said anomalies;
- e) implementing on said simulation engine one or more simulation processes of the real inspection process by varying at least one architecture or a parameter so as to define two or more scenarios (SCEN), said simulation processes such as to provide, for each of said scenarios (SCEN), an output comprising:
- a set of virtual images of said real physical object for each of said scenarios (SCEN);
- one or more architectures of said real acquisition system corresponding to said scenarios (SCEN);
- one or more parameterizations of said real acquisition system corresponding to said scenarios (SCEN);
- f) applying an interpretation model (IM) to said set of virtual images so as to associate to each virtual image, architecture and parameterization corresponding to the scenarios (SCEN) defined in step e) the probability P that the anomaly is represented on the basis of said anomaly model;
- g) selecting one or more virtual images, one or more architectures and one or more parameterizations such that said probability P satisfies the relation P≥TP according to the previous step f), in order to obtain an optimal set of architectures and an optimal set of parametrizations;
- h) choosing, arranging and configuring the components of the real acquisition system according to said architectures and said parameterizations in order to obtain an acquisition system that maximizes the probability of identifying said real anomaly in said real physical object;
- i) by means of the optimized real acquisition system defined in step h), obtaining a set of real images of said object in the real acquisition conditions corresponding to said architectures and to said parameterizations.