GPS And Total Station Combined Positioning Method And Its Application In Precision Agriculture

With the improvement of agricultural machinery operation quality requirements,GPS systems are widely used in agricultural machinery navigation to obtain accurate positioning of agricultural machinery.The obtained GPS position information needs to undergo coordinate transformation to adapt to different needs.In the path planning,data analysis and operation quality evaluation of agricultural machinery navigation,GPS data is often transformed to a Gauss projection plane,but Gauss projection has the characteristics of large coordinate values and inflexibility.During measurement,such as field sampling and mapping,GPS often performs combined measurement with total station.By solving the coordinate transformation model through the control points,the geodetic coordinates of the target points can be obtained.Due to uncontrollable factors in the measurement process,outlier points may be included in control points,decreasing the measurement accuracy.In response to the above problems,this article mainly studies the following aspects:（1）In view of the inflexibility of Gauss projection coordinate system,a local coordinate system establishment method based on farmland shape was proposed.Based on the principle of spatial coordinate transformation,the feasibility of the transformation between geodetic coordinate system and local coordinate system was analyzed theoretically.The steps of establishing local coordinate system were put forward,and the positioning method and path planning method of agricultural machinery with GPS installed in local coordinate system were put forward.（2）In view of the problem that the control points mixed with outliers reduce the measurement accuracy in GPS and total station combined measurement,the method of random sample consensus（RANSAC）algorithm participating in improving the measurement accuracy of GPS and total station combined measurement was studied.After analyzing the research status of GPS and total station combined measurement,the RANSAC algorithm was introduced to improve the coordinate parameter solution process.The direct linear transformation（DLT）and RANSAC algorithm were studied,and the steps of solving the transformation model between geodetic coordinate system and total station coordinate system were given.Besides the algorithm flow chart solved in MATLAB was represented.（3）The tests of the local coordinate system of the farmland were conducted.Firstly,the preciseness of the spatial coordination transformation was verified.The distance of 10fixed intervals were measured with GPS,and then the interval lengths under the Gauss projection plane and the that under the ECEF coordinate were calculated.The result proved that the length of the Gauss projection was shorter than the actual distance,and the error of the ECEF coordinate length was smaller.Secondly,two farmlands in the research base of South China Agricultural University were selected to conduct the local coordinate system establishment tests.The origins of the local coordinate system were determined and the GPS system was used to record the trajectory of the rice transplanter working in the field.The final transformation model was determined by the origin and Euler angle solved by the GPS track.The transplanter’s local coordinates wre obtained then and the lateral deviation of the test trajectory was calculated under the local coordinate.（4）Combined measurement and verification test of GPS and total station was conducted.In this paper,the GPS system and laser total station were used to measure the set target points,and then using DLT combined with the RANSAC algorithm（three threshold levels were set）the transformation model between the two was obtained through the coordinates of the corresponding control point.,and the accuracy of the model is verified with the target points measured by GPS.The test results are:when the threshold was selected as 0.03 meters according to the accuracy of the instrument,there were an error of 0.153×10^-6 degree in longitude,an error of 0.083×10^-6 degree in latitude,and an error of 0.015 meters in elevation.Then the simulation data was used to verify the ability of the method to exclude outliers.The results show that when the proportion of outliers in control points was 25%,50%and 75%,respectively,the errors in the Euler angle and three translation parameters were less than that of the least square method.