We have implemented an eight-antenna module configuration at Jicamarca to perform a three-dimensional imaging of the atmospheric brightness distribution. As an initial test, measurements have been made in the troposphere (5-10 km) and in the equatorial electrojet (EEJ) (95-105 km). Even though the EEJ irregularities are aligned with the magnetic field, i.e., only a two-dimensional approach is needed, we decided to make these observations (1) to calibrate our eight-module configuration, and (2) to study the performance of different imaging methods. These two goals are feasible, given the fact that we have a good knowledge of the EEJ's mean position, orientation, and aspect sensitivity. Four different methods have been implemented to solve for the inverse problem, i.e., to get the atmospheric brightness from the visibility samples on the ground. We used Fourier-based, Capon, and maximum entropy (MaxEnt) methods that were originally used in radar astronomy. In addition, we implemented a fitting technique where a brightness distribution, characterized by a number of anisotropic Gaussian blobs, is assumed. From the EEJ results, (1) there is good agreement between Capon and MaxEnt methods, particularly when the signal-to-noise ratio (SNR) is high, (2) under low SNR, MaxEnt works better than the other techniques, and (3) our fitting technique using two Gaussian blobs seems to work very well, but it is very sensitive to the initial parameters needed to start the fitting procedure. Finally, the tropospheric images indicate that the troposphere over Jicamarca was too homogeneous and no significant gain in information was attained by using more than three-receiving antennas, particularly when long integration times were used (>2 min). The situation could be different in a troposphere less stable than the one at Jicamarca, where the Peruvian coast temperature inversion and stability are prevalent. In the future, efforts will be made to extend the three-dimensional approach to the tropopause and lower stratosphere, where we expect to get more interesting images. © 2001 Elsevier Science Ltd. All rights reserved.