Mapping Indicators of Machinery Utilization Predicted by an Artificial Neural Network

Adrian Aragón-Ramírez, Akira Oida, Hiroshi Nakashima, Juro Miyasaka, Katsuaki Ohdoi


A methodology is presented to generate digital maps containing values of Mechanization Indicators (Mechanization Index and Machinery Energy Ratio), predicted without direct calculation, using a multilayered ANN model. The inputs to the ANN model were simple data obtained from local databases.

Complementarily there were processed digital maps related to parameters on land slope, farm size, soil texture, water supply for crop production and distribution of the land productivity potential for the main crops in the region of study.

Overlapping among the generated maps assisted to analyze the mechanization conditions in every production unit of the Mexican State of Guanajuato, in order to estimate the intensity and suitability of mechanization as well as to identify which farms in the region would benefit more from machinery use.

The developed methodology can facilitate the analysis to prioritize areas for the introduction or replacement of agricultural machinery.

It is concluded that the present methodology would be a good tool to assess mechanization sustainability of agricultural activities; this in turn providing policy-makers and planners with tools with which to judge the best use of land in the near future. Planning the intensity and suitability of mechanization using this approach would contribute to optimize the use of inputs from oil sources.

Full Text: