Abstract
Under natural conditions, light has an impact on cherry fruit recognition, which will affect whether machine vision can accurately identify cherries and then affect the accuracy of machine picking. Through the technical analysis of the convolutional neural network image recognition algorithm, this paper proposes a cherry recognition algorithm that combines morphological filtering and convolutional neural network. The algorithm uses homomorphic filtering to compensate for natural light. Through the natural light compensation method, the problem of low accuracy of cherry recognition caused by light is solved, and then the convolutional neural network algorithm was used to accurate cherry recognition. Compared with a single convolutional neural network algorithm, the use of homomorphic filtering to compensate natural light has good identification under the conditions of front light, backlight and side light, with recognition rate increased by 5%, 20% and 25% respectively. In addition, the running time is shortened by 20%, This paper provid a new method for cherry identification in natural environment.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call