Feature Extraction and Accurate Localization Based on Hough Transform in Complex Scenes

Complex industrial scenes, operation of large machinery or precision machinery, often need to judge, recognition and positioning of industrial scenes, to achieve intelligent systems to improve production efficiency. In view of the shortcomings of traditional manual recognition due to the fluctuation of subjective factors of operators, insufficient reliability, large variance of discnmination accuracy and low efficiency, and the need for a strong applicability system in combination with the variable relative distance between machinery operation and recognition objects, this study proposes a feature extraction and accurate positioning method based on Hough transform in complex scenes. This method reduces the noise interference and edge burr of the image by noise reduction filtering and feature extraction, and completes the feature extraction and accurate localization in complex scenes by introducing multiple relative threshold parameters. Finally, the method is put into work condition discrimination applications and real-time experiments are conducted. The experimental results prove that the Hough transform-based feature extraction and accurate positioning method in complex scenes is widely applicable, real-time, robust, low-cost and stable and reliable.