Least squares method used in measurements adjustment is very sensitive to the occurrence of gross errors and mistakes in the data. In the event of greater number of gross errors, their harmful, imposing the effect, may be as both strengthen and eliminated by which the detection and location are difficult. In order to reduce the number and size of disturbances shall apply procedures of reliability to the stage of measurement, and requires a preliminary analysis of the measurements, generally called preadjustment gross error detection. In aerotriangulation, where there are four groups of measurements with very different origin, the three of them have large systematic errors, which are determined in the adjustment, detection is cumbersome and lengthy, and as practice shows do not always fully effective. Developed method uses two general ways of preadjustment gross errors detection. The first is the detection in several stages, which separates the influence of gross errors and mistakes in the data contained in each group of measurements so that they do not overlap each other. At the stage for this method are only outlier in the measurement of angles of orientation of images. The second is statistical testing of differences between the measured angles and angles of orientation of the images obtained from the adjustment without the use of measurement of the IMU. The method also allows estimation of the standard deviation of IMU measurement. The presented method is only a coarse approximation for more accurate methods. It is assumed that the method detects gross errors with large values. After the detection gross errors and mistakes, observations remaining in the calculations are of course retested during the final standard adjustment. Effectiveness testing of method have been carried out on the 11 aerotriangulatinos large blocks images taken in the country in 20082010. The effectiveness of the method is satisfactory, because undetected outliers are rare. Also, the estimation of standard deviation of measurement of angles is satisfactory.
