Using ANFIS to predict picking position of the fruits sorting system

Abstract

Sorting is one of the important tasks in production line and it has an appreciable effect to the homogeneous of products. Using vision system to increase productivity in automatic sorting system has attracted many researchers. In this paper, a vision sorting system based on computer that can identify the position and properties of fruits has been recommended. The sorting system uses the high resolution camera placed on the top of the body on conveyor belt. With the images are captured by the camera, the software will perform the algorithms to identify the characteristics of the object (for example tomatoes, passion fruits) to sort and give the location of objects to database, then the data of objects can be used by pick and place process that synchronize with controller of the actuators. The sorting process including some steps as detecting the object, determining the object properties like color, size, and shape,…, locating of the object, and using the ability of ANFIS (Adaptive Neuro-Fuzzy Inference System) to calculate the actual gripping position. A five layer neural network of ANFIS is used to adjust input and output parameters of membership function in a fuzzy logic controller. The hybrid learning algorithm is used for training this network. In this algorithm, the least square estimation method is applied for the tuning of linear output membership function parameters and the error hybrid method is used to tune the nonlinear input membership function parameters, it is possible to predict position exactly for actuators and reduce interpolation errors of the system. The algorithms will be experimented on tomatoes and passion fruits, results will be analyzed and evaluated to calculate the stable of the vision system, so that the productivity and accurate efficiency is appropriate.