Pulse Compression technique is a vital tool commonly used in radar to increase range resolution and signal to noise ratio. Pulse compression allows achieving the performance of a shorter pulse using a longer pulse and hence gain of a large spectral bandwidth. Unwanted signals from sidelobes returns affect the detection capability of any radar. Different sidelobe reduction/cancellation techniques based on pulse compression for Linear Frequency Modulated (LFM) radars have been deployed and addressed before. In this paper, a new optimum filter for enhancing radar detection capabilities of LFM radars is introduced. The proposed filter response is compared with the windowed classical matched filter response associated with Hamming window function. The filter is implemented using Software Defined Radio (SDR). A practical test has been carried to investigate its performance. Results show superior performance of our proposed matched filter compared to that of classical versions