Approaches are presented in this work, which contribute to the development of knowledge-based systems, firstly, to assess and improve the quality of production, secondly, to determine the parameters of a control process. These approaches can be classified as a further development of the research in the field of artificial intelligence.The first system controls the evaluation of the quality within a production process using a camera. The colors in the image will be segmented with the help of an image processing tool (black and white or color image). Each segment corresponds to an object or sub-object in the data set (film). The properties of the objects are presented in an object-matrix. These properties characterise quality, which will be evaluated later.The system requires knowledge of the existing objects in order to assign a meaning to these objects in the evaluation process. This knowledge is provided in the form of rules. If knowledge-based systems consist only of rules, they are called "rule-based systems" (Rajendra & Priti, 2010).Since in this work only rules are used, the concepts of knowledge-based system, rule-based system and expert system are synonymous. The strategy of evaluation is based on these rules (rule base) in the following steps:• Filtering of objects• Splitting of connected objects • Cleanup of properties of object• Determination of the number of existing objectsIn this way objects are located and prepared for the evaluation-phase. A reference quality was determined by asking experts. Based on these expert data, the rule-based system evaluates the quality of the objects by comparison with the reference quality and proposes a plan of correction to improve quality.The quality of the production is assessed through the combination of image processing and knowledge-based system. The knowledge-based system is suitable for the use in such existing production systems in which the evaluation of quality is possible. For this purpose, only a new appropriate evaluation strategy (rule interpreter) is required. The performance of the developed system is shown in the first example in the field of quality management. The rule-based system may also be used to determine the parameters of a control process, to improve the quality of productions.First, the parameters of a control process can be measured by several tests. Then, the knowledge of this process is illustrated in rules by machine learning. There are important parameters of a process established to create a predictive system for that. These rule-based system inverse functions for these parameters are implemented.The performance of the developed inverse functions is tested in the second example in this thesis.1