Recently, transfer learning techniques with a base convolutional neural network (CNN) model have widely gained acceptance in early detection and classification of crop diseases to increase agricultural productivity with reducing disease spread. The transfer learning techniques based classifiers generally achieve over 90% of classification accuracy for crop diseases using dataset of crop leaf images (e.g., PlantVillage dataset), but they have ability to classify only the pre-trained diseases. This paper provides with an evaluation scheme on selecting an effective base CNN model for crop disease transfer learning with regard to the accuracy of trained target crops as well as of untrained target crops. First, we present transfer learning models called CDC (crop disease classification) architecture including widely used base (pre-trained) CNN models. We evaluate each performance of seven base CNN models for four untrained crops. The results of performance evaluation show that the DenseNet201 is one of the best base CNN models.

PURPOSES: In this research, an SB3-level roadside barrier for a highway transition zone that meets the newly established guide Installation and Management Guide for Roadside Safety Appurtenance is developed. Its performance is evaluated by a numerical simulation and real-scale vehicle impact testMETHODS: The commercial explicit dynamic software LS-DYNA is utilized for impact simulation. An FE model of a passenger vehicle developed and released by the National Crash Analysis Center (NCAC) at George Washington University and a heavy goods vehicle (HGV) model developed by the TC226/CM-E Work Group are utilized for impact simulation. The original vehicle models were modified to reflect the conditions of test vehicles. The impact positions of the passenger vehicle and truck to the transition guardrail were set as 1/2 and 3/4 of the transition region, respectively, according to the guide.RESULTS : Based on the numerical simulation results of the existing transition barrier, a new structural system with improved performance was suggested. According to the result of a numerical simulation of the suggested structural system, two sets of transition barriers were manufactured and installed for real-scale vehicle impact tests. The tests were performed at a test field for roadside safety hardware of the Korea Highway Corporation Research Institute.CONCLUSIONS: The results of both the real-vehicle impact tests and numerical simulations of the developed transition barrier satisfied the performance criteria, and the results of numerical simulation showed good correlation with the test results.