Research On Auxiliaryalignment Technology For The Vegetable Planting Machinery Working On Ridge Based On Machine Vision

Vegetables are the crops with the largest total sown area in China except food crops.The planting methods of pepper,tomato,Chinese cabbage,broccoli and other vegetables in China are planting on the ridge.The transplanter and planter used to plant the above vegetables needs to follow the ridges.The operator needs to observe the direction of the front ridge and adjust the direction of the planting machinery on the ridge in real time,which puts forward higher requirements for the judgment ability,reaction speed and operation ability of the direction control personnel,and also increases the labor intensity,making it difficult for them to be in a high-efficiency and high-quality working state for a long time.In addition,the operators of some models can not accurately observe the direction of the front ridge when working with their back to the front ridge or their line of sight is blocked.Therefore,it is necessary to provide auxiliary alignment parameters for the control direction personnel of the transplanter through the visual interface,so as to reduce the operation difficulty and labor intensity.Since the existing ridging machine mainly relies on manual control to walk in a straight line,the straightness of the ridged field is difficult to ensure,and the bending of the ridged line often occurs,so the satellite positioning system with a specified path in advance cannot be used to obtain the auxiliary alignment parameters.Machine vision technology collects ridge image information in real time and identifies the alignment path,which is more in line with the operation environment on the ridge.Therefore,this paper studied mechanical aided alignment system of vegetable planting on ridge based on machine vision.The main research work is as follows:(1)This paper analyzed the requirements of the ridge vegetable planting machinery auxiliary alignment system,selected the image acquisition equipment and machine state acquisition equipment,and completed the construction of the auxiliary alignment system.Calibrated the selected camera and determined the installation position of the camera,established the conversion relationship from the pixel coordinate system to the machine coordinate system,and predicted the region of interest in combination with the forward speed of the machine and the running time of the algorithm.The visual interface of the auxiliary alignment system was designed,and the dynamic trajectory model was established based on the vegetable planting chassis on the ridge.(2)The recognition method of ridge line based on color information was studied.The color characteristics of ridge and furrow in the ridge color image and the principle of color difference were analyzed.The result showed that the gray difference between ridge and furrow could be increased by gray reconstruction using brightness component.An autonomous threshold segmentation method was proposed to calculate the threshold of segmenting the current gray image in real time,and the contour of the target ridge was extracted from the binary image by using the contour detection method.Finally,based on the target ridge contour,an approximate quadrilateral method was proposed to extract the ridge center line as the alignment path of the ridge vegetable planting machine.(3)The recognition method of field ridge line based on depth information was studied.The depth image was transformed into the height image by coordinate system transformation.A sub region Otsu method was proposed to segment the height image,and the maximum connected region extraction method was used to extract the target ridge region in the ridge image.Finally,the row traversal method was used to mark the characteristic points of the ridge boundary,and the ridge boundary line was fitted based on the least square method.The linear equation of the ridge center line was calculated according to the ridge boundary equation,and the ridge center line was used as the alignment path of the ridge vegetable planting machine.(4)The row regulation strategy of vegetable planting machinery on ridge was studied.Through the coordinate system transformation,the position of the alignment line path in the world coordinate system was obtained,and the position relationship between the alignment path and the axis of the machine in the world coordinate system was analyzed.The steering directions corresponding to different position relationships were obtained,and the theoretical steering radius and steering time were calculated.Finally,the alignment control strategy proposed in this paper was simulated and verified.The simulation results showed that the maximum deviation distances for the ridge of 0.4m,0.8m and 1.2m were 35.6mm,49.8mm and64.2mm respectively,and the alignment accuracy was centimeter level,which met the actual row requirements of the planting machinery on the ridge.(5)Auxiliary alignment system experimental research.Ridge line recognition was the basis of the auxiliary alignment system.Therefore,this paper carried out experiments in different ridge environments to verify the effectiveness of the ridge line recognition method.For the ridge line recognition method based on color information,light color,dark color and film covered ridge were used for experiment,and for the ridge line recognition method based on depth information,0.08 m,0.15 m and 0.25 m high ridges were used for experiment.The experimental resulted showed that the recognition success rate of the two methods was more than 95%,the average running time of the algorithms was less than 0.3s,and tthe average deviation angle Δθ was less than 1°,and the average deviation distance Δx were less than 9mm.In order to compare the two ridge line recognition methods,the two methods were used to recognize the same ridge image.The recognition results showed that the ridge line recognition method based on depth information was better than the ridge line recognition method based on color information.At last,the method of recognizing ridge and row line based on depth information was applied to the field experiment.The experimental results showed that the row regulation suggestions provided by the visual interface could guide the manual control of the machine to achieve the purpose of accurate row alignment.