Search

CN-121999465-A - Pond transplanting ridge-seeking navigation line identification method based on circularity and edge pixel point jump

CN121999465ACN 121999465 ACN121999465 ACN 121999465ACN-121999465-A

Abstract

The invention discloses a recognition method of a ridge-finding guidance route for transplanting after a pond based on circularity and edge pixel point jump, and belongs to the technical field of ridge farming visual recognition. The method comprises the steps of image acquisition, image boundary smoothing, color conversion, threshold segmentation, denoising, edge detection, midpoint extraction and navigation line fitting, wherein a depth camera is used for acquiring a depth image, a traditional RGB image is converted into an HSV image to reduce the influence of illumination conditions on furrow identification, H components are used for carrying out next processing, threshold segmentation is used for acquiring a binarized image, a morphological denoising method and an area threshold method are used for removing most noise interference in the image, depth denoising of the image is realized based on circularity calculation, meanwhile edge pixel point jump is used for extracting a ridge pit boundary, then horizontal coordinates of two boundary points of the ridge pit are used for obtaining midpoint coordinate points of image pixels, and finally the point coordinates are used for realizing the navigation line fitting of the ridge pit by using a least square method.

Inventors

  • WANG FENG
  • Ding Yongmeng
  • LI KUNZHEN

Assignees

  • 西南林业大学

Dates

Publication Date
20260508
Application Date
20260108

Claims (8)

  1. 1. The method for identifying the ridge-seeking navigation line after pond transplanting based on circularity and edge pixel point jumping is characterized by comprising the following steps: s1, performing image acquisition on ridge pits through a depth camera, and acquiring RGB color images and depth images; s2, converting the depth image from an RGB color space to an HSV color space through a conversion model based on the RGB color image, and outputting an HSV color image; s3, dividing and constructing an H component target image according to the H component based on the HSV color map; S4, based on the H component target image, threshold segmentation is carried out to obtain a binary image; S5, morphological denoising is carried out through closing operation and opening operation based on the binary image, and the denoised binary image is output; The closing operation is expansion and corrosion; the open operation is performed by corrosion and then expansion; S6, calculating the area and the circularity of each communication area based on the denoised binary image, and screening out ridge pit areas according to the area threshold value and the circularity threshold value to obtain a binary image of the ridge pit areas; s7, extracting left and right edge points based on a binary image of the ridge pit area by using an edge pixel point hopping algorithm, and respectively fitting by adopting a least square method to obtain a left boundary straight line and a right boundary straight line; And S8, calculating midpoint coordinates point by point according to the left boundary straight line and the right boundary straight line, and then fitting a straight line to all midpoint points by using a least square method to obtain a navigation line, and completing the recognition method of the ridge-seeking navigation line after the pond in which the circularity and the edge pixel point jump.
  2. 2. The recognition method of post-pool transplanting ridge-seeking navigation line based on circularity and edge pixel point jump according to claim 1, wherein in S1, the image acquisition is specifically: Based on the self-propelled auxiliary transplanting platform, shooting video of a working scene, decomposing the video into images frame by frame, and establishing a coordinate system by taking the upper left corner of the image as an origin of coordinates (0, 0), wherein the ground clearance of a depth camera during shooting is 20-100 cm, the inclination angle is vertically downward 90 degrees, the shooting frame rate is 30 fps, and the resolution is 1920 multiplied by 1080.
  3. 3. The method for identifying post-pool transplanting ridge navigation lines based on circularity and edge pixel point jump according to claim 1, wherein in S2, the expression of the conversion model is as follows: where H is hue, S is saturation, V is brightness, r is red intensity value, g is green intensity value, and b is blue intensity value.
  4. 4. The method for identifying post-pool transplanting ridge navigation lines based on circularity and edge pixel point jump according to claim 1, wherein in S4, the threshold segmentation adopts a fixed threshold method, wherein the threshold of the H component is 20 at the lower limit and 120 at the upper limit.
  5. 5. The recognition method of the post-pool transplanting ridge-seeking navigation line based on circularity and edge pixel point jump according to claim 1, wherein in the step S5, morphological denoising is performed through a closed operation and an open operation, the closed operation is performed by first expanding and then corroding, the open operation is performed by first corroding and then expanding, the size of a processing kernel K is set to 10 multiplied by 10 pixels, and an expansion formula expression is as follows: the corrosion formula is expressed as follows: in the formula, Is the relative coordinate offset inside the kernel.
  6. 6. The method for identifying post-pool transplanting ridge navigation lines based on circularity and edge pixel point jump according to claim 1, wherein in S6, the circularity threshold is set to 0.6, and the area threshold is set to 10000 pixels.
  7. 7. The method for identifying a post-pool transplanting ridge guidance route based on circularity and edge pixel point hopping according to claim 1, wherein in S7, specifically comprising: (1) Establishing a left matrix L and a right matrix R for storing the coordinates of the jump points; (2) Carrying out progressive scanning on the binary image of the ridge pit area: when the pixel value jumps from 0 to 1, the position coordinate is recorded and stored in a left matrix L; when the pixel value jumps from 1 to 0, the position coordinate is recorded and stored in a right matrix R; (3) And respectively adopting least square fitting to the points in the left matrix L and the right matrix R to obtain a left boundary straight line and a right boundary straight line.
  8. 8. The method for identifying post-pool transplanting ridge navigation lines based on circularity and edge pixel point jump according to claim 1, wherein in S8, the formula for calculating the coordinates of the center point by point is: Wherein, the Is the left boundary straight-line abscissa of the graph, Is the right boundary straight-line abscissa of the graph, As the vertical ordinate of the left boundary, Is the right boundary straight line ordinate.

Description

Pond transplanting ridge-seeking navigation line identification method based on circularity and edge pixel point jump Technical Field The invention relates to the technical field of ridge culture agricultural vision recognition, in particular to a recognition method of a post-pool transplanting ridge-seeking navigation line based on circularity and edge pixel point jump. Background The ridge culture crops are one of the main cultivated crops in China, and relate to a plurality of varieties, wide planting area and high yield, especially the climate in the south is rainy, a large number of vegetable plants are planted in a ridging cultivation mode, and the main machine is configured into a single-ridge or double-ridge ridger, a double-row or four-row transplanter and a self-propelled auxiliary harvesting platform. With the continuous improvement of the agricultural mechanical level in China, the demand of agricultural machinery operation from 'organic availability' to 'organic good use' is increasingly strong, the agricultural machinery operation is required to develop towards an automation and unmanned direction, and a visual recognition technology is one of basic guarantee technologies of the development direction. The method is characterized in that the agricultural working scene is identified based on the machine vision technology, required information is finally extracted from an image through the steps of preprocessing image information, target detection, target tracking and the like, and then measurement and judgment are carried out instead of human eyes, and the autonomous high-precision navigation of the ridge culture crop agricultural machinery equipment is realized by combining the technologies of tracking control and the like, so that the method becomes a problem to be solved in the current ridge culture crop vision identification technology. Currently, a machine vision-based ridge culture navigation method has been studied to a certain extent, and mainly focused on extracting crop row or furrow characteristics by using a traditional image processing algorithm (such as edge detection, hough transformation and the like). For example, there are methods to extract a guidance line through color segmentation of RGB images in combination with hough line detection, and also studies to introduce a depth camera to reduce illumination effects. However, these existing technologies still have obvious disadvantages when applied to the field of ridge transplanting, such as difficulty in identifying due to sensitivity to illumination and soil background and irregular shape of seedling pits, difficulty in considering both real-time and precision, and insufficient adaptability, so how to quickly, accurately and robustly identify seedling pits on ridges and extract navigation reference lines in complex field environments becomes a key problem to be solved urgently for improving autonomous operation precision of ridge transplanting machines. The problem is solved, so that the automatic transplanting operation pushing device is beneficial to pushing automation and unmanned operation, and has important significance for improving the operation efficiency, reducing the labor intensity and realizing refined agriculture. Disclosure of Invention In order to solve the problems, the invention provides a recognition method of a ridge-finding navigation line after pond transplanting based on circularity and edge pixel point jump. In order to achieve the purpose, the invention provides the following technical scheme that the recognition method of the post-pool transplanting ridge-seeking navigation line based on circularity and edge pixel point jump specifically comprises the following steps: s1, performing image acquisition on ridge pits through a depth camera, and acquiring RGB color images and depth images; The invention is based on the self-propelled auxiliary transplanting platform to collect the image of the video section shot in the operation scene, decompose the collected video into images frame by frame and establish a coordinate system, take the upper left corner of the image as the origin (0, 0), take the right corner coordinate of the top of the image as the width,0, and take the left corner coordinate of the bottom of the image as the (0, height), collect RGB color images and depth images. S2, converting the depth image from the RGB color space to the HSV color space through a conversion model based on the RGB color image, and outputting an HSV color map. And S3, dividing and constructing an H component target image according to the H component based on the HSV color diagram. S4, based on the H component target image, threshold segmentation is carried out to obtain a binary image; The expression for threshold segmentation is: in the formula, Is an H-component target image. S5, morphological denoising is carried out through closing operation and opening operation based on the binary image, and the denoised binary im