Depending on the size of storage space or limited transmission capacity compression of sensor data is sometimes needed. There exist non-destructive (lossless) compression methods but lossy compression methods are more efficient. These lossy methods will distort the sensor data. One way to examine how this distortion affects automatic signal processing methods is to test them in a simulation environment. This project has investigated how compression can be used in the simulation environment MSSLab. In addition, a study has been made on how an automatic signal processing method handles sensor data that has been subjected to lossy compression. A lossy data compression methods (JPEG) has been added to MSSLab and its impact on a detector have been investigated. Whether the detector must be trained on the compressed images or on the original images in order to achieve the best possible outcome has been investigated. Based on images with seven different degrees of compression, a number of ROC curves have been developed to investigate the detection algorithm. No or low compression often produces good detection results, but sometimes a heavily compressed image can give even better results. It seems that the compression when acts as a pre-filtering that highlights properties that the detector can benefit from. Hence, it is difficult to clearly say how the degree of compression affects the selected detection algorithm.