Autonomous cis-lunar spacecraft navigation is critical to mission success as communication to ground stations and access to GPS signals could be lost. However, if the satellite has a camera of sufficient quality, line of sight (unit vector) measurements can be made to known Lunar Landmarks (e.g. Tycho Crater) to provide observations which enable autonomous estimation of position and velocity of the spacecraft, that can be telemetered to those interested space based or ground based consumers. An improved Batch Gaussian-Initial Orbit Determination (IOD) DC algorithm, based on the exact values of the f and g series (free of the 8th order polynomial and range guessing), for spacecraft state estimation, is mentioned here and applied in the inertial coordinate frame (2-Body Problem) to initialize a (non-conic based) CRTBP Extended Kalman Filter (EKF) navigator that collects angle only measurements to a known location on the Moon to sequentially estimate position and velocity of an observer spacecraft flying on an approximate southern L1 Halo orbit. It was found that the best approach is to initialize the CRTBP EKF (Navigator) using the solution from the (2-body inertial) f and g series Batch DC Filter with at least ten measurements taken against a LEO satellite expressed in the 2-body inertial frame. Thereafter it is best to continue the Navigator with subsequent measurements taken against the center coordinates of the Tycho crater (expressed in the CRTBP rotating frame). For successful conic-based Batch Filter initialization and long term CRTBP EKF convergence, it was found that the cadence for all optical measurements should be no more than ten minutes, for a simulated measurement noise of 0.1 ? one sigma uncertainty about the line of sight measurement unit vector.