We propose a simple scheme for computing the locations of restoration sensors for a disconnected WSN to recover the network's connectivity to a certain level, while keeping the number of restoration nodes as few as possible. Since an accurate, optimal solution to the problem is NP-hard, we resort to tackling the problem in a greedy and heuristic manner. We just deploy enough new nodes to establish k disjoint paths between disconnected components, where k is the network's original connectivity. That is, the scheme just grants the minimally necessary condition for restoring k-connectivity. The extremely time-consuming task of making sure the result is indeed a k-connected network will not be carried out. Empirical study has been conducted to evaluate the effectiveness of the proposed approach.