Thermographic cameras are becoming increasingly popular in all kinds of diagnostic testing aiming to assess the technical state of thermal power engineering machinery, equipment and installations. The cameras provide thermograms that enable identification of various irregularities, including thermal bridges. However, it can be observed that their accuracy and reliability depend substantially on the conditions in which tests are carried out. Industrial thermographic testing, especially in thermal power engineering, is one of the most difficult procedures. Apart from the advantages of this particular technique, this paper presents an analysis of the effect of the conditions in which tests are performed on the obtained results. Attention is drawn to parameters characterizing the environment and the tested element surface, such as emissivity, reflected temperature, distance between the camera and the object, air temperature, etc. The sources of errors in the testing and in the interpretation of thermograms are indicated. Methods are also presented that enable elimination of irregularities, which improves the accuracy of the final results.