In this paper, we study simultaneous multi-channel spectrum sensing in cognitive radio context. The sensing receiver RF front-end is assumed to deploy wideband multi-channel IQ down-conversion which is well-suited for highly-integrated circuit implementations. Such RF front-end is, however, also prone to several RF impairments, such as IQ imbalance, which leads to mirror-frequency cross-talk and can thus considerably degrade the sensing performance. Assuming that the individual channel sensing is building on energy detection, we first analyze the spectrum sensing performance of the multichannel sensing receiver in terms of false alarm and detection probabilities. The analysis shows that IQ imbalance is especially harmful in terms of false alarms stemming from the mirror-channel crosstalk, and is greatly emphasized when the overall down-converted signal has high dynamic range. Motivated by this, we present an efficient digital compensation scheme, called energy correction method, where the sensing statistics of any particular channel are properly adjusted based on the corresponding mirror-channel statistics such that the effect of the crosstalk is minimized. Optimum minimum mean-squared error (MMSE) solution for the energy correction processing is first derived, complemented then with a practical low-complexity sample estimator. Extensive computer simulations demonstrate that under various parameter settings for IQ imbalance and receiver dynamic range, the proposed scheme yields sensing performance practically identical to the IQ imbalance free reference performance. Thus the proposed method offers feasible RF impairment-aware energy detection solution for practical multichannel sensing receivers.