Evaluation Of Land Consolidation Potential Based On Neuralnetwork Method

For a long time, China is faced with a large number of people and a large number of extensive land area, but a small number of arable land, less resources per capita arable land, shortage of arable land resources reserve. With the development of economy, the phenomenon of wasting and destroying in land use is becoming more and more serious, which makes the contradiction between population and land rather sharp, so how to stick to 18 million mu of arable land red line, to protect arable land, to fully tap the potential of land use become important problems needed to be addressed. Land consolidation and fully tap the potential of land use can effectively increase the available land area, ease the contradiction between population and land, improve farmers’ production and living conditions, and promote the coordinated development of people and land, promote the pace of agricultural modernization and new rural construction. Land consolidation potential evaluation is one of the preliminary work of land consolidation, and it provides some basis for the reasonable arrangement of land consolidation projects, the smooth development of land consolidation.This paper combines the evaluation of land consolidation potential with BP neural network to construct an evaluation model which uses BP neural network to evaluate land consolidation potential. In the paper land Consolidation potential is divided into rural residential land consolidation potential and rural land consolidation potential, and respectively from the natural, economic, social and ecological four aspects to build the evaluation index system. By referencing to various materials, the potential levels of rural residential land consolidation and rural land consolidation were established, and divided into five levels, namely higher potential Ⅰ, high potential Ⅱ, medium potentialⅡ, low potential and lower Ⅱpotential Ⅴ. Take 50 townships in Dacheng, Jinghai, Huanghua, Botou which get close to Qingxian as the training sample. The neural network can automatically obtain its reasonable rules between parameters. A case about Qingxian, evaluation units for 11 regions, include 6 towns, 4 townships, 1 farm. Then input the actual values of the 11 regions to the network,combine Matlab with the actual values to obtain the most close to the standard value of the corresponding evaluation level. The evaluation results are as follows: For rural residential land consolidation, Grade Ⅰpotential area includes Qingzhou, Machang and Jinniu town, these areas belongs to the higher potential areas; Grade Ⅱpotential area includes Pangu township, Liuhe town and Caosi township, these areas belongs to the high potential areas; Grade Ⅱ potential area includes Mumadian town and Shangwu township, land consolidation potential in the medium level and land consolidation potential is moderate; Grade potential Ⅱarea includes Xinxing town, Chenzui township and the farm, these areas belongs to the low potential areas. For rural land consolidation, Grade Ⅰ potential area includes Qingzhou and Machang town, these areas belongs to the higher potential areas; Grade Ⅱ potential area includes Jinniu town, Pangu township and the farm, these areas belongs to the high potential areas; Grade Ⅲ potential area includes Liuhe town,Shangwu township and Caosi township, land consolidation potential in the medium level and land consolidation potential is moderate; Grade Ⅳ potential area includes Xinxing town, Chenzui township, these areas belongs to the low potential areas.