Method and test system for assessing the quality of a multi-channel micro- and/or subwavelength-optical projection unit

A method for assessing the quality of a projection unit being at least one of a multi-channel micro-optical projection unit and a multi-channel subwavelength-optical projection unit, comprising the following steps:

• illuminating at least a predefined portion of the projection unit so that an image is generated by at least two channels of the predefined portion of the projection unit;
• capturing the image generated by the predefined portion of the projection unit;
• analyzing the image;
• determining at least one characteristic quantity based on the analysis of the image, wherein a value of the characteristic quantity is related to at least one of at least one characteristic feature of the projection unit, at least one defect of the projection unit and at least one defect class of the projection unit;
• and
• assessing the quality of the projection unit based on the at least one characteristic quantity,
• wherein the image of the projection unit is analyzed via at least one of a statistical method and via a machine learning module,
• wherein the machine learning module comprises a pre-trained artificial neural network,
• wherein the machine learning module has been pre-trained with labelled training data,
• wherein the labelled training data comprises sample images generated by the predefined portion of a sample optical projection unit, and wherein the labelled training data comprises at least one of the at least one characteristic quantity corresponding to the respective sample images and the quality class of the respective sample projection unit, and
• wherein the at least one characteristic quantity corresponding to the respective sample images is obtained by applying the statistical method to the sample images.