This work addresses the challenge of providing precise runway-relative position, velocity and orientation reference to a landing aircraft based on monocular camera and low-cost inertial sensor data complemented by barometric sensor readings. GPS information is excluded from this sensor set because it is intended to use the developed estimator for GPS and Instrumental Landing System fault detection. The characteristic size of the runway is assumed to be unknown and it is estimated run-time after verifying global observability of the nonlinear system. The delay caused by image processing is dealt with a delayed-Error-State Kalman Filter (ESKF). This algorithm considers dynamic propagation of the image information between acquisition and application forward in time thus the delay does not appear in the system dynamics. The first evaluation of the estimator is done for ideal simulated data to verify applicability of the delayed-ESKF and its flawless implementation. Then more realistic simulated data with sensor biases and noise is considered to verify closer to realistic performance and bias estimation precision. Finally, the estimator is tested with real flight data collected in the VISION EU H2020 research project. Estimation results are compared to GPS SBAS-based data and Airbus precision tolerances showing satisfactory performance. The methodological contribution of the paper is the unique combination of existing methods and ideas leading to a new solution proven to work satisfactorily even with real flight data.