Navel orange damage detection method

Abstract

The invention relates to the field of navel orange detection, in particular to a navel orange damage detection method. The method is used for solving the problems of unstable and inaccurate detection results and the like in traditional navel orange detection. The method comprises the following steps: acquiring a navel orange image, and preprocessing the image; extracting surface color features, surface defect features and size features of the navel oranges; establishing a high-quality navel orange sample library and an inferior navel orange sample library, and training by using a BP neural network to obtain a trained navel orange damage detection model; converting the characteristic value into a characteristic vector, inputting the characteristic vector into a trained navel orange damage detection model, and carrying out navel orange detection; and outputting a final detection result to realize navel orange damage detection. According to the method, the collected navel orange image is preprocessed, namely denoising and background segmentation are carried out, the noise influence of the navel orange image is removed, and a plurality of threshold values are set, so that the navel orange is separated from the background more perfectly; the navel orange damage detection model is obtained through BP neural network training, so that the detection efficiency of the model is greatly improved.

Description

A kind of method for damage detection of navel orange technical field The present invention relates to navel orange detection field, relate in particular to a kind of method for navel orange damage detection. Background technique In the traditional navel orange detection, generally through the navel orange sorting machine, the fruit is rotated on the assembly line, and then a plurality of Cameras or line scan cameras capture all or most of the surface information of the fruit. However, this method of image acquisition is easy to The fruit causes mechanical damage; the serial acquisition method is used for the fruit image, and the acquisition speed is slow; and because of the poor shape of the fruit different, so that the posture on the assembly line has a certain randomness, which increases the difficulty of subsequent processing, which is not conducive to the later analysis of the navel orange map. Image processing, which in turn affects the extraction of navel ora...

AI Technical Summary

Problems solved by technology

[0004] Therefore, the object of the present invention is to provide a method for navel orange damage detection, in order to solve the problems such as unstable and inaccurate detection results in traditional navel orange detection

Images