A new technique is described that can be used to estimate basic parameters such as target size (projected area), surface temperature, and spectral emissivity based on the optical radiation measured by a multiband infrared radiometer. This is of great interest in science and industry where the ability to measure temperature and surface optical characteristics via noncontact methods is vital to the analysis and control of many processes. This new technique is designed to compensate for reflected radiation from other background sources not in the sensor field of view. The technique is based on applying a simple closed form expression for the optical signal measured by the sensor and expanding the unknown emissivity in a series of orthogonal functions. To obtain the parameters of interest nonlinear least-squares estimation techniques are employed. To test the overall procedure data were collected on targets specifically designed to span the range from gray to highly nongray. The optical radiation from these targets was measured in a ground-based space simulation chamber in the presence of known and controlled warm background sources. The targets were instrumented with thermocouples that measured the surface temperature throughout the experiment. Results of the data analysis show that, using a three- or four-band radiometer, it is possible to measure surface temperature to a high degree of accuracy, and that the technique is robust in the presence of noise. Other applications are also explored. These include the simultaneous estimation of the target size (projected area) for point targets at a known distance from the optical sensor, as well as a detailed estimate of the material spectral emissivity using data provided by a spectrometer.