Road surface garbage sensing method for intelligent road sweeping

Abstract

The invention discloses a road surface garbage sensing method for intelligent road sweeping. The method comprises the steps: establishing and marking a garbage image database; using a data enhancementmethod which comprises geometric transformation and color transformation of the image; randomly scaling, cutting and arranging the images; expanding a data domain by utilizing a generative adversarial network; positioning and recognizing large garbage and small garbage existing on the road surface by adopting target detection and density estimation combined sensing; after a rectangular frame andlabels of the garbage are obtained through target detection, converting the rectangular frame into a density map form, and assigning different density weights according to different labels; combiningthe density map obtained by conversion with a density map generated by a density estimation algorithm to obtain a final pavement garbage density image; calculating candidate cleaning points, and inputting the candidate cleaning points into a path planning module; based on the obtained garbage distribution information, inputting the garbage distribution information the path planning module, and adjusting the driving path. Intelligent sweeping of the sweeper is achieved, and high practical value is achieved.

Description

technical field [0001] The invention relates to the field of artificial intelligence, in particular to a road garbage perception method for intelligent road cleaning. Background technique [0002] A: Zhang Pengcheng. A method for urban street garbage detection and cleanliness assessment that integrates mobile edge computing and deep learning. The Faster R-CNN target detection network is used to detect road garbage. Detectable garbage categories include: waste paper, plastic bags, plastic bottles, cans, etc. However, it cannot effectively detect dense small garbage such as leaves, branches, and cigarette butts. And there are few training samples, and only 321 garbage images are used as the training data set, which cannot fully learn the network model. [0003] B: Mohammad Saeed Rad. A Computer Vision System to Localize and Classify Wastes on the Streets. A deep convolution-based target detection network is used to detect and identify small garbage on the road, including l...

AI Technical Summary

Problems solved by technology

It has a good effect on sparse small targets, but it cannot be effectively perceived when the targets are highly overlapping

Moreover, the algorithm has high requirements on the image resolution. When the image resolution is low, it cannot be effectively detected, and the high-resolution image will greatly affect the real-time performance of the algorithm.

At the same time, in order to eliminate the target scale change caused by the perspective transformation of the camera, the imaging plane must be parallel to the ground when the camera is installed, which also brings a certain degree of complexity to the garbage detection work.

[0004] Since the image information occupied by small garbage is less, and the appearance features change greatly after a high degree of overlap, the conventional target detection model adopted by A and B cannot effectively detect dense small garbage, and the garbage data used for model training is less, the model cannot be fully studied, which is also not conducive to the detection of road litter

Images