QUALITY ANALYSIS AND GRADING OF SOYBEAN USING MACHINE LEARNING
Keywords:ANN, Grading, Image Processing, Machine Learning, Soybean, Morphological, Seed features
The use of good quality seed is very important for the better production of a good quality crop and is essential for export in markets. Quality control process is very important in food industry, based on quality of food products are classified and graded into different grades. Soybean is primarily graded based on its grain shape, colour, size and texture. This paper attempts to automate the grading process by using image processing and machine learning techniques. Soybean’s grade is affected by damaging, decolourization, infection by insects, immaturity and shrivels, splitting, breaking, cracks, inorganic and organic foreign matter present in the sample. One of the objectives of this paper is to study the effect of these parameters on shape, colour, size and texture of the soybean image. In the present soybean-handling scenario, type and quality are identified manually by visual inspection which is tedious and not accurate. There is need for the growth of accurate, fast and objective system for quality determination of food grains. This paper is automate the system for grading of soybean by extracting morphological features as attributes for classification using image processing techniques and artificial neural network. The classification is done on the basis
of features extracted from the segmented images. Simple averaging is used for the combination of results from each classifier. The data is divided into the training and testing data. The data classified as training data is then used for training the neural network. Remaining data is used for the testing purpose. The training and testing dataset involves both normal as well as abnormal cases. Then classification is done into normal and abnormal. It provides better results compared to single neural network that has accuracy of about 90%. This method requires minimum time and it is low in cost.
How to Cite
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.