The System Design Of Green Crop Area Detection Based On Monocular Vision Technology

The growth state parameters of the field crops under natural conditions, such as covering area, density and height are important for predicting the yield and assessing quality, and have become one of the important research tasks in ecological agriculture information monitoring.This paper mainly studied green crop covering area acquisition, including two aspects:1) counting pixels of target crop image in the image coordinates;2) converting pixels in the image coordinates to actual crop growing area through the coordinate transformations, The methods adopted in this paper contained crop image preprocessing,8-neighborhood-boundary tracking and counting target pixels with CRgn class of MFC. The internal and external parameters of the camera were acquired by the camera calibration procedures, Then, green crop growth covering area in the world coordinate system was obtained through integrating the image coordinate system, plane coordinate system, camera coordinate system and world coordinate system.In this paper, considering the difficulty of error analysis of field crop area, the calibration plate of Shaanxi Weishi was used to implement the experiment procedures about errors analysis, where the crop area acquisitions under the different heights and angles of camera set in the field conditions were simulated in order to analysize errors and control error regions of the whole image. The experimental results showed that the system designed in this paper was feasible, the error with X-direction:1.84mm, and Y-direction:11.89mm, predicting the area of celery:0.964m2、5.835m2、0.29m2、0.675m2、0.12m2、0.35m2.0.38m2、2.432m2、0.37m2、0.13m2, with the error rang from0.5%to6.9%.