The performance of image coding techniques can be enhanced through the utilization of a priori knowledge. Critical features of the image are first identified and then treated more favorably in the encoding process. In aerial surveillance imagery, thin lines and point objects constitute critical features of interest whose preservation in the encoding process is important. For the human visual system, coding degradation at low rates is more detrimental for these features than for the edges that constitute boundaries between regions of different contrast. A highly nonlinear, matched-filter-based algorithm to detect such features has been developed. Pre-enhancement (highlighting) of the detected features within the image prior to coding is shown to noticeably reduce the severity of the coding degradation. A yet more robust approach is to pre-enhance the slightly smoothed image, giving rise to an image in which all critical thin lines and point objects are crisp and well defined at the cost of nonessential edges in the image being slightly rounded off. For the transform coding techniques, distortion parameter readjustment and variable-block size coding are promising alternatives to the pre-enhancement approaches. In the former, subblocks containing any part of the detected critical features are kept within a low distortion bound by means of the local rate adjustment mechanism. The latter approach is similar except that the image is partitioned into various size subblocks based on the extracted feature map.