| Crop spraying is the important link in the process of agricultural production. In order to realize the efficient utilization of spraying potions, key technology of target spraying rabot for crops in field was studied, based on cauliflower Transplanted. This paper put forward a method of crops and crop row information acquisition, designed a control method for row-follow based on autonomous transverse motion as auxiliary in artificial driving, and target spraying on crops based on machine vision. On this basis, a target spraying rabot system for crops in field was developed, tested and analysed. The main research contents and results are as follows:(1) A method for recognizing and locating crops based on colour and crops distribution characteristics was studied. After a background segmentation way of (G-R)&&(G-B), the pixel histogram was obtained by projection in row or column direction. The peak and trough of the pixel histogram was segmented by the Otsu method, and then searched for peak area boundary, which is the crop or crop row area boundary, from the pixel histogram above the Otsu threshold. According to the transplanting characteristics of equal crop row, a filter method based on maximum matching degree with crop row was designed, making the num of crop row detecting and real crop row equal, and guaranteeing the correctness of following image processing. The results of crop recognizing test for series frame images showed that average time consuming is18ms, average crrect recognization rate is94.93%, average error recognization rate is5.07%.(2) A method for ceterline extraction of crop row based on least square method was studied. Each crop area boundary was recognized and located by projection-searching. The average of wo boundary in the same direction of each crop area is the crop center coordinate. Taking all the crops center belonged to the middle crop row as the data points, fitting straight line by least square method is crop row centreline. This method is high accuracy, and less time consuming.(3) A method for row-follow based on autonomous transverse motion as auxiliary artificial driving was studied. The tractor steering information guided working personnel to steer tractor to adjust the tractor’s pose, and meanwhile eliminate large row-follow deviation. The autonomous transverse motion based on machine vision further eliminated row-follow deviation. The3tests in order was carried out, including "artificial driving along crop row","tractor steering information as guide for auxiliary artificial driving along crop row", and"tractor steering information and autonomous transverse motion as auxiliary in artificial driving along crop row". The results showed that the average of crop row offset distances in order were42mm,30mm,20mm, when the forward speed was0.161m/s.(4) A method for target spraying on crops based on machine vision was studied. The distance between nozzle and crop, and crop canopy width detected by machine vision, integrating speed detection and timing real-time determined relative position of nozzle and crop area boundary. The nozzle was controled to open when nozzle approaching to the crop, and close when nozzle leaving the crop, realizing crops target spraying. The target spraying test results showed that the effective spraying rate is85.88%, the error spraying rate is11.76%, when the forward speed was0.609m/s. (5) Target spraying robot system for crops in the field was developed, including machine vision system, transverse mechanism, spraying waterway, control system, power source and so on. Multistage power structure with main gasoline generator was designed for power supplied to control system. One hydraulic output was transformed into two hydraulic outputs for hydraulic power source on front three-point suspension and transverse mechanism. The results of integration test of row-follow and target spraying showed that the effective spraying rate is81.18%, the error spraying rate is15.29%, when the forward speed was0.609m/s. |
PDF Full Text Request