Distance Measurement Method based on Two-Dimensional Image Recognition and Inverse Perspective Transformation

摘要

At present, 80% of the information people receive every day comes from the vision. The traditional direct contact measurement method cannot meet the highly intelligent requirements in operation and manipulation. In the current research, a novel method in measuring the distance was proposed to simplify the steps and promote the efficiency in terms of the two-dimensional image recognition and inverse perspective transformation. This new method was developed by primarily establishing the prior knowledge base and the auto-correlation function MATLAB program of Harris corner detection. Then, the image processing was implemented algorithmically by using the inverse lens transform method. The results showed that this method could measure the distance between two objects contained in an image with the allowable errors. It was showed that this method had high stability, high reliability and high adaptability.