Target detection method based on semantic Hough transformation and partial least squares

Abstract

A target detection method based on semantic Hough transform and partial least square method, involving computer vision technology. First, use the training set to establish a regression model between the feature vector of the image block, the semantic feature vector, the category of the image block and the voting offset vector, and then substitute the feature vector and semantic feature vector of each image block in the image to be detected into the regression model , find out the image blocks with positive categories and generate votes to form a Hough image, and finally find the local maximum point of the voting density in the Hough image as the detection result to complete the target detection.

Description

technical field [0001] The invention relates to computer vision technology, in particular to a target detection method based on semantic Hough transform and partial least square method. Background technique [0002] Computer vision was originally separated from the field of digital image processing in the 1960s and became an independent research direction, and has been widely used in aerospace, automatic navigation, industrial inspection, medical research, clinical diagnosis and treatment, security monitoring and tracking, national defense , transportation, remote sensing and many other important fields. Among them, object detection, especially human-related detection, is a key technology closely related to various practical applications. Take public transportation, which urban residents must participate in every day, as an example. More than one million people die in traffic accidents around the world every year, but accurate and fast pedestrian detection technology can de...

AI Technical Summary

Problems solved by technology

When the number of thresholds and parameters is larger, it is more difficult to choose the appropriate value, and it is difficult to produce the best effect

Images