Image transmission over HF communication is suitable for long-distance coverage beyond the line of sight and across intercontinental. Unfortunately, the channel condition is degraded with low Signal to Noise Ratio (SNR) and multi-path spreading. In this paper, we aim to transmit images over HF channel and introduce some techniques to improve the received noisy images. First, a Least square estimation technique is applied to estimate the channel state information, then we propose an enhanced lter to reduce image noise eect and improve the received image appearance. This proposed enhanced lter is based on our enhanced version of the Median Filter (MF) and the Mathematical Morphology (MM). Simulation results assess the performance of the proposed lter compared with the performance of some popular lters, we conclude that the proposed enhanced lter can be used for any further detection and recognition operations, and it introduces promising results which can be used for real-time implementation with dierent computer vision applications.