Self-adaptive automobile fuel tank outer cover identification method based on regional contrast difference

Abstract

The invention discloses a self-adaptive automobile fuel tank outer cover identification method based on region contrast difference. The method comprises the following steps: carrying out self-adaptive histogram equalization on an initial image to obtain an outer cover region contrast-enhanced image; performing high-order Gaussian blur on the image to expand a highlight area; carrying out difference on the image before and after blurring to obtain a color feature image; graying the color feature image and then carrying out threshold segmentation to obtain a feature region image; obtaining circumscribed rectangles of all the regions, and performing preliminary screening on three basic conditions, namely areas, length-width ratios and region positions, of all the regions based on morphological characteristics of the shape of the outer cover; performing secondary screening on suspected areas obtained by preliminary screening by adopting different types of outer cover contour judgment methods based on outer cover morphological characteristics; and adaptive identification of outer cover areas of different types and sizes is realized based on generality of morphological characteristics. The method has the advantages of high adaptability, high environmental stability, high recognition precision and the like.

Description

technical field [0001] The invention belongs to the technical field of computer vision, and in particular relates to an adaptive automobile fuel tank cover recognition method based on area contrast and difference. Background technique [0002] With the increase in the use of automobiles in our country, automobile refueling operations have become a common and important service. The traditional way of refueling is to manually complete the refueling process. Due to the danger of the gas station environment and the remote location of some gas stations, manual refueling operations have problems such as high labor costs and insufficient work efficiency. With the application and promotion of intelligent technology, the construction of unattended robot gas stations has become a hot topic of research. In the process of robot intelligent refueling, the recognition of the outer cover of the automobile fuel tank is the premise and basis for completing subsequent actions. [0003] The a...

AI Technical Summary

Problems solved by technology

The document "Robotic fueling system" (Andrews M, Thompson C, Warner E, et al.IEEE International Conference on Technologies for Practical Robot Applications.2008) proposes a RFID system identification and positioning method combined with a distance sensor, using the distance sensor to obtain the position of the vehicle object Information, using RFID to identify the position of the fuel tank cover, but this identification method needs to install RFID tags on the car, which is not practical enough; the literature "Monocular, vision based, autonomous fueling system" (Farag A, Dizdarevic E, Eid A, et al. Application of Computer Vision. IEEE, 2002) proposed a fuel tank cap recognition method based on monocular vision, using the form of pasting logo pictures on the fuel tank cover to assist in the recognition of the fuel tank cap position, this method has a high Positioning accuracy, but still need to add auxiliary objects, the process is too limited to achieve practical application and promotion; Chinese invention patent CN201910507058.8 discloses a vision-based refueling port target recognition method, which performs threshold segmentation on the equalized image , using contour template matching to get the position of the fuel tank cover, but this method needs to filter out the highlighted part and realize the position recognition in the form of prior knowledge (fixed template type) during feature extraction. Realize adaptive recognition

[0004] In summary, the existing research cannot meet the recognition requirements of fuel tank cap objects of different oil types and sizes without adding auxiliary objects.

Images