Affected by noise, light blocking, color and shape of object, the quality of captured stripes in structured light three dimensional (3D) measurement system is degenerated. As the quality of captured stripes is one of the key factors for measurement accuracy, some large error data is introduced into the measurement results which can only be recognized artificially with prior knowledge of the object to be measured. In this paper, a method is proposed to estimate the quality of stripe image. In the method, two parameters, skewness coefficient of stripe gray distribution and the noise level, are used to estimate the quality of stripe. The simulation results show that the bigger the skewness coefficient is, the bigger the error of stripe locating results is. Meanwhile, the smaller the noise level is, the smaller the error of stripe locating results is. The method has been used to estimate the experimental image, and the same conclusion can be obtained. The method can be used for recognizing large error data automatically by the two parameters.