Smartphone assisting convolutional neural networks for soil texture classification in dry and wet humid conditions in West Guwahati, Assam

Soil texture using a hydrometer or pipette method requires expertise, although these are accurate. A soil expert may help the farmer to detect by analyzing visual of soil, which is not always This paper presents smartphone image-based sand and clay classification in wet dry humid conditions Self Convolution Neural Network (SCNN) finetuned MobileNet.A dataset 576 images was prepared low-cost under natural light conditions. Different augmentation techniques such as shift, range, rotation, zoom were applied increase number dataset. The best performance MobileNet reported at epoch 15 with testing training loss 0.0091 0.0194, respectively. Though SCNN model performed 10 accuracy 99.85%, less computation time (167.8s) than (273.2s). precision recall models 99.62 (MobileNet) 99.84 (SCNN). itself model, whereas computing different can be used replicate traditional analysis farmers use it for better productivity.