Scenic spot ground plastic bottle detection method

Abstract

The scenic spot ground plastic bottle detection method comprises the following steps: 1) collecting images of high-altitude cameras in a large number of scenic spots and data sets of other plastic bottles, calibrating the data sets according to on-site management requirements, and determining a used one-stage target detection algorithm model; 2) constructing a parameter self-adaptive loss function; 3) constructing a loss function LOSS of a one-stage target detection algorithm model; 4) updating the weight of the one-stage target detection algorithm model by adopting a gradient descent method until the model converges; and detecting the plastic bottles by using the trained model in an actual system, and achieving the detection system for the plastic bottles according to the existing time and position relationship of the plastic bottles on site. The method has the advantages that the proposed focus loss function can improve the parameter adaptability of the plastic bottle detection model, and the accuracy of plastic bottle detection is greatly improved.

Description

technical field [0001] The invention belongs to the technical field of image recognition and computer vision, and relates to a method for detecting plastic bottles on the ground of scenic spots. Background technique [0002] At present, in view of the fact that tourists discard plastic bottles at will in the scenic area, the staff cannot deal with the problem of discarded plastic bottles on the ground in time. Identify the plastic bottles on the ground of the scenic spot. Among them, the processing of plastic bottles on the ground of the scenic spot is completed through the inspection of the staff. This processing method requires a lot of manpower, material resources and financial resources, and due to the fact that the staff needs to take vacations and there will be missed inspections during the manual inspection process, etc. , the effect is not ideal. However, the generalization of traditional image algorithms is poor, and the camera needs to be at a fixed angle and und...

AI Technical Summary

Problems solved by technology

Although the two-stage target detection model has better detection accuracy, its forward reasoning speed is slow and cannot meet the real-time requirements of business scenarios

In the traditional one-stage target detection algorithm model, although the real-time performance of the algorithm is better, it cannot reach the detection accuracy of the two-stage target detection algorithm model

Th...

Images