Features extraction in agricultural products using computer image processing

Abstract

The study was based on the application of image processing and machine vision technology in the extraction of some features of interest in agricultural products. It involved the development of an algorithm that used some computer programs written in the FORmular TRANslation (FORTRAN 95) programming language to carry out tests on some input data images. The input data images were obtained by digitizing the original images of some Amaranthus vegetables, cowpea seedlings and some food samples on which some fungi were found to be growing as captured by a digital camera. The results of this study carried out on the images showed that the algorithm was able to extract both the boundary of the leaves and the crop position by discriminating between the crop and the uncovered soil surface as well as the growth of the fungi on the food samples. The extracted images as produced by the programs compared favourably well with the original images as captured by the digital camera. This visual comparison showed that the approach is very promising in developing automated sorting and quality inspection techniques which can be applied to some agricultural and food products.