Weights of indirect measurements in inverse photogrammetry tasks

Abstract

This is the fourth article in the series of publications on ways of improving the statistical quality of solving inverse photogrammetry tasks. It deals with researching the impact of indirect measurement weights on the equalization’s statistical quality. The author established that following its criteria is the key to the most plausible distribution of errors on the effects of indirect measurements. Since the accuracy of the equalization is estimated by the remainder residuals of the indirect measurements groups, the reliability of these evaluations depends on how strictly the criteria for the statistical quality of the equalization are met. The standard of indicators’ accuracy reliability is obtained. An algorithm for adjusting their weights during equalization was developed. An experimental approbation of the technique was performed. Based on the results of the calculations, the effectiveness of the joint application of all ways developed in this and previous articles is established to improve the statistical quality and obtain a reliable estimate of the equalization computations accuracy. In addition, it turned out that the application of the developed algorithms enables increasing the numerical stability of the mentioned calculations and ensure the stability of the results. The expediency of clarifying the initial values of direct measurements of the equalized parameters is confirmed. The explanation of the obtained results in terms of covariance analysis is given.