Experimental Study on Instrument Pointer Detection Based on Hough Transform and RANSAC Algorithm
Minglei Zhao, Hongbin Yu, Hongyu Shao · 2021 4th International Conference on Algorithms, Computing and Artificial Intelligence · 2021
Because of the complex working environment in most factory workshops, the measurement of pointer instrument relies on manual reading. At present, in the field of instrument detection, most algorithms are based on improved image preprocessing and ignore the accuracy, stability and robustness of automatic pointer selection. In order to improve the anti-noise performance of meter pointer selection, RANSAC (random sampling consistent) algorithm was combined with Hough transform linear detection, and the meter pointer in the binary edge graph was straight-line fitting. The idea of improvement is to use SURF algorithm to extract feature points first and separate points distributed near different lines. Further, RANSAC algorithm is used to assume the linear model data, given a reasonable inner group, and remove the invalid outlier data. Finally, the Hough transform algorithm is used to fit the instrument pointer line. The linear fitting calibration image verification of the instrument detection pointer shows that the combined algorithm has better fitting effect on the instrument detection pointer, reduces a lot of noise interference and improves the accuracy.