<![CDATA[Honey is a food substance that has sweet taste and thick structure produced by bees. Honey can be distinguished by observing the color and clarity of honey, but it is quite difficult due to lack of knowledge and each type of honey has almost the same color and level of clarity. Based on these problems, this study designed a system to classify the type and authenticity of honey. The sensors used are the TCS3200 sensor, the LDR sensor, and the pH sensor. The TCS3200 sensor and LDR sensor are placed on the back side and on the front side of the sample glass an LED light is added, while the pH sensor is at the top of the glass and the pH sensor eye is immersed in the solution in an upright position so that the sensor can optimally determine each characteristic of the honey sample. The backpropagation ANN algorithm in this study is processed using Arduino Nano with a network structure of 3 inputs, 1 hidden layer containing 24 perceptrons, and 1 output which is divided into 6 classes. The structure design process uses 900 datasets, the learning rate is 0.001, the epochs are 28.451 and the training process is 2 hours 23 minutes 14 seconds. From the testing process the backpropagation neural network algorithm is proven to be able to classify each honey class well and the NN algorithm accuracy reaches 94.45%, with an average computation time of 0.80076 seconds.]]>
Copyrights © 2022
