CN101794391B - Greenhouse environment leading line extraction method - Google Patents

Abstract

The invention provides a greenhouse environment leading line extraction method which is characterized by comprising the following steps: S1, collecting greenhouse road surface information as original colored image to be transmitted into a computer; S2, grayscaling the original colored image, and obtaining gray level image; S3, carrying out image segmentation on the gray level image, removing plant interference in a greenhouse, and realizing road surface extraction; S4, after image segmentation, taking the obtained binary image as a mask to be covered on the original colored image, and obtaining image after overlapping the mask; and S5, processing the image after overlapping the mask by uniformization according to R, G and B components, and extracting a leading line. The method carries out uniformization on a RGB colored space which is processed by simple mathematical translation, and the RGB colored space can be directly acquired from a camera without complicated transition so as to have good arithmetic real-time property. Furthermore, the method is suitable for extracting the leading line in the greenhouse under different illumination conditions, and is stable and reliable in effect.

Description

A kind of greenhouse environment leading line extraction method

Technical field

The invention belongs to field of agricultural robots, relate under a kind of adaptation different illumination conditions walking robot leading line rapid extraction technology in the greenhouse.

Background technology

Airmanship is one of core technology of autonomous mobile robot, and the mobile robot has only pose, the walking path information of knowing self exactly, could carry out autokinetic movement safely and effectively.Because the complexity of agricultural robot working environment changes and unpredictability, increased the navigation difficulty of agricultural robot greatly.Vision guided navigation technology as one of comparatively ripe at present agricultural robot airmanship, be meant and utilize vision system to identify under the agricultural environment mark information as walking path, relative position according to agricultural robot and walking path calculates controlled quentity controlled variable, by a series of processes that turn to topworks to regulate its position and then follow the tracks of expected path.Accurately quick identification is extracted the gordian technique that mark information is the agricultural robot navigation.

Mark information comprises natural landmark and artificial two kinds of forms of road sign, and the research of agricultural robot navigation at present by the identification to physical feature in the working environment, realizes navigation mainly based on natural landmark.But natural landmark features such as plant row or ridge are complicated and changeable under the greenhouse, based on the navigation information poor reliability of vision extraction.Therefore study artificial landmark navigation mode in the greenhouse, significant for the development and the practical application of agricultural robot.

Summary of the invention

(1) technical matters that will solve

In the acquisition process of vision guided navigation image, often be subjected to the interference of various factors, the particularly ambient light interference of (comprising intensity and color) under the greenhouse, the leading line extraction algorithm of cutting apart especially of navigation picture is formulated and brought very big difficulty.The present invention is directed to artificial navigation road sign in the greenhouse, adopt colour TV camera to gather navigation route information, propose the leading line extraction algorithm that a kind of substep is removed disturbing factor, satisfy the requirement of real-time and robustness.

(2) technical scheme

In order to solve the problems of the technologies described above, the present invention has designed a kind of greenhouse environment leading line extraction method, it is characterized in that, said method comprising the steps of:

S1: hand paste colourful navigation line on the greenhouse road, gather the greenhouse information of road surface, be sent to computing machine as original color image;

S2: described original color image is carried out gray processing handle, obtain gray level image;

S3: described gray level image is carried out image segmentation, remove the greenhouse implants and disturb, realize the road surface extraction;

S4: the bianry image that obtains after the image segmentation is covered on the described original color image as masking-out, obtain the image after masking-out superposes;

S5: the image after the described masking-out stack is carried out normalized according to R, G, B component, extract leading line.

Wherein, the gray processing formula that is adopted among the described step S2 is:

grayimage＝0.30×R+0.59×G+0.11×B

Wherein R, G, B are respectively original color image rgb color passage gray-scale value.

Wherein, the image partition method that is adopted among the described step S3 is the constant automatic threshold method of square.

Wherein, the formula of normalization RGB color space is among the described step S5:

$\left\{\begin{array}{c}r=\frac{R}{R+G+B}\\ g=\frac{G}{R+G+B}\\ b=\frac{B}{R+G+B}\end{array}\right.$

Wherein r, g, b are respectively the component value in the color space after the normalization.

(3) beneficial effect

In technique scheme, conversion can obtain through simple mathematical by the RGB color space in normalization RGB color space, and the RGB color space can directly obtain from video camera, does not need complicated conversion, and the algorithm real-time is good.And the present invention can adapt to leading line extraction under the different illumination conditions in the greenhouse, effect stability, reliable.

Description of drawings

The green cucumber plant original image that Fig. 1-a is in the embodiment of the invention to be adopted;

Image after the green cucumber plant original image gray processing that Fig. 1-b is in the embodiment of the invention to be adopted is handled;

Fig. 1-c is the image of green cucumber plant gray level image after threshold value is separated processing in the embodiment of the invention;

Fig. 1-d is that the image after the green cucumber plant Threshold Segmentation covers the image that obtains on the original image in the embodiment of the invention as masking-out;

Fig. 1-e is a leading line extraction effect synoptic diagram in the embodiment of the invention;

Fig. 2 is the process flow diagram that extracts leading line in the embodiment of the invention.

Embodiment

Below in conjunction with drawings and Examples, the specific embodiment of the present invention is described in further detail.Following examples are used to illustrate the present invention, but are not used for limiting the scope of the invention.

The feature of natural landmark such as plant row or ridge is complicated and changeable under the greenhouse, causes the inefficacy of landmark navigation system easily.In order to improve the reliability of navigational system, the present invention is hand paste redness leading line on the greenhouse road, and the navigation vision system is discerned this leading line, and then calculates controlled variable, realizes stable navigation tracing task.

The present invention has disposed the monocular vision acquisition system for realizing the vision guided navigation task, is used to gather the greenhouse information of road surface, extracts leading line.The EVI-D100P colored CCD monopod video camera that this vision collecting system adopts Sony company to produce, the camera acquisition simulating signal collects in the computing machine by the Osprey-200 video frequency collection card.

The present invention has used a kind of substep to remove the leading line extraction algorithm of disturbing factor.

1) image gray processing is handled

(Fig. 1-a) carries out gray processing to be handled, and obtains gray level image (Fig. 1-b) to original color image.The gray processing formula is:

grayimage＝0.30×R+0.59×G+0.11×B

Wherein R, G, B are respectively original color image rgb color passage gray-scale value.

2) automatic threshold method image segmentation

The present invention adopts the constant automatic threshold method of square to 1) obtain gray level image and tentatively cut apart, remove plant and disturb, realize that the road surface extracts, effect is shown in Fig. 1-c.

3) bianry image after the Threshold Segmentation is covered on the original color image as masking-out, shown in Fig. 1-d.Partly be capped (pixel brightness value is 255) corresponding to the bianry image white portion in the coloured image, all the other black part divide the maintenance original pixel value constant.After this step operation, both removed plant to extracting the interference of red leading line, kept the leading line information of redness in the original image again, be convenient to the carrying out of postorder operation.

4) red leading line is extracted

To R, G, B component by formula (1) carry out normalized, obtain r, g, b studies as characteristic parameter.

$\left\{\begin{array}{c}r=\frac{R}{R+G+B}\\ g=\frac{G}{R+G+B}\\ b=\frac{B}{R+G+B}\end{array}\right.---\left(1\right)$

For the cucumber green plant, g chromatic component value is greater than r component and b component; Cement pavement and red leading line r chromatic component value are respectively greater than g chromatic component and b chromatic component.Yet the difference (r-g) of red leading line r chromatic component value and g chromatic component will be much larger than the difference (r-g) of cement pavement r chromatic component and g chromatic component.The threshold value of setting r-g can be cut apart red leading line and cement pavement well greater than 0.15.Utilize the process flow diagram of r, g under the normalization RGB color space, the red leading line of b component extraction, as shown in Figure 2.Cut apart by flow process shown in Figure 2, the pixel on the red leading line is shown as white, other pixel is set to the black background look, and resulting effect is shown in Fig. 1-e.

The above only is a preferred implementation of the present invention; should be pointed out that for those skilled in the art, under the prerequisite that does not break away from the technology of the present invention principle; can also make some improvement and replacement, these improvement and replacement also should be considered as protection scope of the present invention.

Claims (4)

1. a greenhouse environment leading line extraction method is characterized in that, said method comprising the steps of:

S1: hand paste colourful navigation line on the greenhouse road, gather the greenhouse information of road surface, be sent to computing machine as original color image;

S2: described original color image is carried out gray processing handle, obtain gray level image;

S3: described gray level image is carried out image segmentation, remove the greenhouse implants and disturb, realize the road surface extraction;

S4: the bianry image that obtains after the image segmentation is covered on the described original color image as masking-out, obtain the image after masking-out superposes;

S5: the image after the described masking-out stack is carried out normalized according to R, G, B component, extract leading line.

2. greenhouse environment leading line extraction method as claimed in claim 1 is characterized in that, the gray processing formula that is adopted among the described step S2 is:

grayimage＝0.30×R+0.59×G+0.11×B

Wherein R, G, B are respectively original color image rgb color passage gray-scale value.

3. greenhouse environment leading line extraction method as claimed in claim 1 is characterized in that, the image partition method that is adopted among the described step S3 is the constant automatic threshold method of square.

4. greenhouse environment leading line extraction method as claimed in claim 1 is characterized in that, the formula of normalization RGB color space is among the described step S5:

$\left\{\begin{array}{c}r=\frac{R}{R+G+B}\\ g=\frac{G}{R+G+B}\\ b=\frac{B}{R+G+B}\end{array}\right.$

Wherein r, g, b are respectively the component value in the color space after the normalization.

Images