Non-intrusive visualization of the structure of flames can offer us many advantages in studying the reaction mechanisms of combustion and observing special distributions of the parameters required for the development of equipment such as jet engines and gas turbines. Process tomography is a relatively new technique for such a task, but is useful owing to its fast speed and capability of detecting signals related to ionizations caused by chemical reactions and thermal effects. Electric capacitance tomography (ECT) is one of the process tomographic techniques. ECT usually comprises a sensor array of electrodes that detect permittivity variations in the measuring zone, a data-logging device and a computer that controls data acquisition and carries out image reconstruction. There have been studies on ECT imaging of flames; however, ECT has not been exploited sufficiently to reveal the inner structure of the flames. In this study, a sensor with planar electrodes is created, and the associated three-dimensional sensitivity map is generated by the finite-element method to detect flame structure. A series of experiments are carried out covering a range of feed rates of fuel and air. Data are collected by the ECT sensor and hardware. The results of the ECT reconstruction show good agreement with actual features, and the structure of the flame is found. This opens up a new route for the study of flames.This article is part of the themed issue ‘Supersensing through industrial process tomography’.