Accurate Feature Detection of Abdominal Pads for Thermal Image Inspection

An abdominal pad used in a low-frequency stimulator transfers electric pulses to the abdomen and waist. Because an abdominal pad is heated to increase the amount of pain relief; however, low-temperature burns can be caused by prolonged wear. Thus, an accurate thermal inspection method for better safety of an abdominal pad is discussed in this paper. The key aspect of this inspection method is an automatic feature detection approach for measuring the thermal characteristics. The shape of an abdominal pad consists of complicated curves, and thus the background temperature is accurately separated using the feature detection method. The features were detected by synthesizing multiple images and dilating a binarized image. The proposed method was implemented using OpenCV and other open sources. The results show that feature detection is possible regardless of the size, position, orientation, and shape of the abdominal pad. Feature detection also improves the measurement accuracy for thermal imaging.