CN102831398A - Tree apple recognition method based on depth image - Google Patents
Tree apple recognition method based on depth image Download PDFInfo
- Publication number
- CN102831398A CN102831398A CN2012102580515A CN201210258051A CN102831398A CN 102831398 A CN102831398 A CN 102831398A CN 2012102580515 A CN2012102580515 A CN 2012102580515A CN 201210258051 A CN201210258051 A CN 201210258051A CN 102831398 A CN102831398 A CN 102831398A
- Authority
- CN
- China
- Prior art keywords
- fruit
- image
- area
- tree
- zone
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Abstract
The invention discloses a tree apple recognition method based on a depth image in the technical field of image recognition. According to the technical scheme, the tree apple recognition method comprises the steps of: scanning an apple tree by using an apple picking robot laser vision system and generating a corresponding depth image; realizing layer-by-layer segmentation of a target region by combining with change characteristics of depth information from the center to the edge of the fruit in the image to obtain one or more binary images; and calculating the circularity of each communicating region in the one or more binary images and the area ratio of the areas of the communicating region to the maximum area, and through restraining the relative shape parameter, accurately recognizing the fruit.
Description
Technical field
The invention belongs to the image recognition technology field, relate in particular to a kind of based on apple recognition methods on the tree of depth image.
Background technology
At present, the main force effect of Tendency of Agricultural Machinery in Chinaization in agricultural production highlights day by day, effectively ensured the progress and the quality of agricultural production.Wherein, agricultural robot is a highest intelligent agricultural machinery, its appearance and be applied as modern agricultural development and opened up wide prospect more.The development of agricultural robot technology will help improving the intelligent equipment of Tendency of Agricultural Machinery in China, improve the present agricultural operation pattern of China.
Apple is one of the world's four big fruit, and results are plucked and accounted for 40% of whole production workload, and the quality of plucking quality directly has influence on storage, processing and the sale of apple, thereby finally influences the market price and economic benefit.The harvesting work of China's apple at present mainly relies on artificial the completion, because the fruit tree plant is higher, the apple maturity stage is shorter, has strengthened the labour intensity of manual picking; The aging of the propelling of Development of China's Urbanization and Chinese population in addition, the rural laborer reduces in a large number, causes the apple-picking manpower shortage at last.The application of apple picking robot will be one of important channel that addresses the above problem.
Machine vision is the maximum external environmental information source of picking robot, not only concerns it and fast, accurately discerns the ability of fruit, has also directly determined the reliability of picking robot, has made number of research projects in related fields both at home and abroad.Existing vision system is many to be key component with the ccd video camera; There are polytrope, non-design feature such as intellectual, opening not owing to pluck environment; Particularly the problems such as uncertainty of illumination condition make the discrimination of fruit and bearing accuracy receive certain limitation.
Apple picking robot laser vision system is through the scanning fruit tree, obtains a large amount of range data, and converts range data into gray-scale value, thereby obtain the system of depth image.Because the range measurement principle of laser vision system; The influence that the depth image that generates is not changed by light; And the pixel in the image has been represented a relative depth information; Therefore, compare with common handled two-dimension optical image, depth image can be resolved shape, size and the far and near relation of different objects in the object scene more easily.
Summary of the invention
The present invention is directed in the existing vision system of mentioning in the above-mentioned background technology, the discrimination of apple and bearing accuracy aspect exist not enough, have proposed a kind of based on apple recognition methods on the tree of depth image.
A kind of based on apple recognition methods on the tree of depth image, its characteristic comprises the following steps:
Step 1: use apple picking robot laser vision system scan fruit tree, and generate corresponding depth image;
Step 2: improve the visual effect of the image that obtains through step 1, suppress garbage, and the image after improving is carried out Filtering Processing;
Step 3: represent the fruit respective peaks of diverse location to search in the gray scale frequency curve chart of the image that step 2 is obtained; And the variation characteristics of fruit center depth information in the combining image to the edge; Realization is successively cut apart the target area, obtains the one or more bianry image;
Step 4: the area ratio of the circularity of each connected region of one or more bianry image that obtains in the calculation procedure 3 and each connected region area and maximum area, through above-mentioned relative shape parameter is retrained, accomplish accurate identification to the fruit zone.
In the step 1, use apple picking robot laser vision system scan fruit tree, obtain a large amount of fruit tree range data, through type (1) converts the range data that collects into gray-scale value, generates corresponding depth image.
In the formula, x, y represent the parameter of both direction, and MinD, MaxD, D (x, y), (x y) is lowest distance value, maximum range value, point (x, distance value y) and gray-scale value respectively to G.
In the step 2,, suppress garbage, adopt the selectivity contrast enhancement process in order to improve the visual effect of depth image.Because the gray-scale value that image histogram occurs belongs to background in a big way, directly ignores respective pixel, will contain that [a, b] expands between the gray area of important information, through type (2) is promptly accomplished the conversion of tonal range by [a, b] to [0,255].In order to take into account smooth noise and to keep details two aspects, adopted median filtering technology, the pixel window through 3 * 3 carries out Filtering Processing to image.
In the step 3, show that the gray-scale value in the depth image is different owing to be in the object of diverse location.Through the gray scale frequency curve map analysis that comprises the fruit image is found that the gray scale frequency curve chart of fruit image does not always present bimodal form, therefore, conventional threshold segmentation method is inapplicable.The present invention proposes a kind of adaptive thresholding value detection method.Concrete grammar is following:
A) set gray threshold Gt, only consider that (x y)≤gray scale change of frequency in the Gt scope, promptly only discerns the top layer apple to G ';
B) based on a) said scope, utilize the evaluation function of setting, search for the peak dot position that all satisfy condition through evaluation function formula (3), and reject the peak dot (non-apple respective peaks) of its peak value less than maximal value 75%; If still show as the multimodal state,, further confirm the validity of peak dot through setting a peak separation threshold value Dt.As D≤Dt, explain that different fruit alternate position spikes distances are less, a back peak dot is confirmed as active position; Work as D>Dt, different fruit alternate position spikes are described apart from bigger, need two peak dots are all kept;
In the formula, the frequency that P (i) occurs for i gray-scale value among the figure, P
MaxBe the maximal value of P (i) neighborhood Ps, Ps={P (i-1), P (i-2) ... P (i-n), P (i+1), P (i+2) ... P (i+n) }, n gets 3 ~ 12 integer.As V (i)=1, expression i gray-scale value is the peak dot position, and vice versa.
C) consider that fruit is spherical, centre of surface has certain depth difference to the edge, and there is some difference to make gray-scale value after the conversion, in order to carry out preserving more complete fruit edge after the binary conversion treatment, with peak dot position S
iAdd one and set side-play amount u as segmentation threshold.Be single peak dot if detect, and then between cut section [0, S
1+ u] comprise the fruit zone; If detect to a plurality of peak dots, can exist between a plurality of cut sections, [0, S
1+ u] expression ground floor fruit zone, [S
1+ (S
2-S
1)/2, S
2+ u] expression second layer fruit zone, and the like.Pixel value in the cut zone is given logical zero, and the pixel value outside the cut zone is given logical one.If the gray scale frequency curve chart of fruit image is unimodal situation, then obtain a width of cloth bianry image; If the gray scale frequency curve chart of fruit image is the multimodal situation, then obtain several bianry images; I=1 wherein, 2 ..., n; N represents the sum of peak dot.
In the step 4, after step 3 obtains the one or more bianry image and negates, applied mathematics morphology disposal route open operation, i.e. corrosion after expansion earlier, the hole in the blank map is removed the burr among the figure; Utilize zone marker with different connected domains among the difference figure, calculate respectively each regional circularity R (j) and with the area of maximum area than A (j), wherein j representes the mark of connected domain; Through scope restriction, with A (j) to above characteristic parameter<1/>5A
MaxOr R (j)<0.6 area pixel assignment 0, other pixels remain unchanged, A
MaxRepresent the area maximal value of connected region; Realization is to the accurate extraction in fruit zone, and will obtain several as a result figure carry out exclusive disjunction, to obtain final complete fruit area image.
Effect of the present invention is:
1) depth image does not receive the influence that light changes, and is easy to from image, resolve the space geometry characteristic of fruit, branches and leaves and mutual hierarchical relationship, can be fruit identification the more pattern information of horn of plenty is provided.
2) the present invention adopts a kind of selectivity contrast enhancement process, can improve the visual effect of depth image, and suppresses garbage, is convenient to the Target Recognition in later stage.
3) the present invention proposes a kind of adaptive thresholding value detection method; Through representing the search of the fruit respective peaks of diverse location in the right gray scale frequency curve chart; The variation characteristics of fruit center depth information in the appended drawings picture to the edge; Obtain rational segmentation threshold, to realize successively cutting apart to the target area.
4) the present invention adopts based on the restriction on the parameters method of area ratio with circularity, to realize the accurate extraction to the fruit zone.
Description of drawings
Fig. 1 is a kind of process flow diagram based on apple recognition methods on the tree of depth image provided by the invention;
Fig. 2 is a kind of depth image based on apple on the tree that collects of apple recognition methods on the tree of depth image that the embodiment of the invention provides; Wherein, a is two depth images that apple is close; B is the depth image of two apple apart from each others;
Fig. 3 is a kind of depth of smoothness image that strengthens based on the contrast of apple recognition methods on the tree of depth image that the embodiment of the invention provides; Wherein, a is two depth of smoothness images that apple is close; B is the depth of smoothness image of two apple apart from each others;
Fig. 4 is a kind of gray scale frequency curve chart based on apple recognition methods on the tree of depth image that the embodiment of the invention provides; Wherein, a is two gray scale frequency curve charts that apple is close; B is the gray scale frequency curve chart of two apple apart from each others;
Fig. 5 is a kind of bianry image based on apple recognition methods on the tree of depth image that the embodiment of the invention provides; Wherein, a is the binary conversion treatment to Fig. 3 a; B successively cuts apart one to Fig. 3 b; C successively cuts apart two to Fig. 3 b;
Fig. 6 is a kind of zone marker image of handling based on the mathematical morphology of apple recognition methods on the tree of depth image that the embodiment of the invention provides; Wherein, a handles the morphology of Fig. 5 a; B handles the morphology of Fig. 5 b; C handles the morphology of Fig. 5 c;
Fig. 7 is a kind of apple zone of extracting based on the form parameter leash law of apple recognition methods on the tree of depth image that the embodiment of the invention provides.Wherein, a is the fruit zone among Fig. 3 a; B is the fruit zone among Fig. 3 b.
Embodiment
Below in conjunction with accompanying drawing, preferred embodiment is elaborated.Should be emphasized that following explanation only is exemplary, rather than in order to limit scope of the present invention and application thereof.
The depth image of embodiment of the invention apple on the tree of gathering is Benq's apple recognition methods on the tree of depth image the example.Concrete scheme is following:
Step 1: use apple picking robot laser vision system scan fruit tree, obtain a large amount of fruit tree range data, through type (1) converts range data into gray-scale value, generates corresponding depth image.As shown in Figure 2.
Step 2: in order to improve the visual effect of depth image, suppress garbage, adopt the selectivity contrast enhancement process.Because the gray-scale value that image histogram occurs belongs to background in a big way, directly ignores respective pixel, will contain that [a, b] expands between the gray area of important information, through type (2) is accomplished the conversion of tonal range by [a, b] to [0,255].In order to take into account smooth noise and to keep details two aspects, adopted median filtering technology, the pixel window through 3 * 3 carries out Filtering Processing to image.The value that contains in the present embodiment between the gray area of important information is [0,76].As shown in Figure 3.
Step T3, the object that is in diverse location shows that the gray-scale value in the depth image is different.Through gray scale frequency curve chart (like Fig. 4) analysis that comprises the apple image is found that the gray scale frequency curve chart of apple image does not always present bimodal form, therefore conventional threshold segmentation method is inapplicable.The present invention proposes a kind of new adaptive thresholding value detection method, reached desirable intelligent segmentation effect.Concrete steps are following:
1. according to the generation characteristics of depth image, set gray threshold Gt, only consider that (x y)≤gray scale change of frequency in the Gt scope, promptly only discerns the top layer apple to G '.Gt=150 in the instance.
2. utilize the evaluation function of setting, search for the peak dot position that all satisfy condition through evaluation function formula (3), and reject the peak dot (non-apple respective peaks) of its peak value less than maximal value 75%.N=8 in the instance.
3. if still show as the multimodal state,, further confirm the validity of peak dot through a peak separation threshold value Dt.As D≤Dt, explain that different apple alternate position spikes distances are less, a back peak dot is confirmed as active position; Work as D>Dt, different apple alternate position spikes are described apart from bigger, need two peak dots are all kept.Dt=15 in the instance.
4. consider that apple is spherical, centre of surface has certain depth difference to the edge, and there is some difference to make gray-scale value after the conversion, in order to carry out preserving more complete apple edge after the binary conversion treatment, with peak dot position S
iAdd one and set side-play amount u as segmentation threshold.U=7 in the instance.
5. if detect and to be single peak dot, and then between cut section [0, S
1+ u] comprise the apple zone; If detect to a plurality of peak dots, can exist between a plurality of cut sections, [0, S
1+ u] expression ground floor apple zone, [S
1+ (S
2-S
1)/2, S
2+ u] expression second layer apple zone, and the like.Pixel value in the cut zone is given logical zero, and the pixel value outside the cut zone is given logical one.Image segmentation result is as shown in Figure 5.
Step T4: behind step T3,, can obtain several bianry images under the multimodal distribution situation, all have the apple zone in every width of cloth to obtaining a width of cloth bianry image under the unimodal situation.Image after cutting apart has some holes and burr, and its back of negating through the morphology open operation, is promptly corroded after expansion earlier, to eliminate above-mentioned interference of noise.Utilize zone marker with different connected domains among the difference figure, as shown in Figure 6; Calculate respectively each regional circularity R (j) and with the area of maximum area than A (j), wherein j representes the mark of connected domain; Through scope restriction, with A (j) to above characteristic parameter<1/>5A
MaxOr R (j)<0.6 area pixel assignment 0, other pixels remain unchanged, A
MaxRepresent the area maximal value of connected region; Realization is to the accurate extraction in apple zone.To obtain several as a result figure (like 5b and 5c) carry out exclusive disjunction, to obtain final complete apple area image.As shown in Figure 7.
The above; Be merely the preferable embodiment of the present invention, but protection scope of the present invention is not limited thereto, any technician who is familiar with the present technique field is within the technical scope that the present invention discloses; The variation that can expect easily or replacement all should be encompassed within protection scope of the present invention.Therefore, protection scope of the present invention should be as the criterion with the protection domain of claim.
Claims (5)
1. one kind based on fruit recognition methods on the tree of depth image, and its characteristic comprises the following steps:
Step 1: use apple picking robot laser vision system scan fruit tree, and generate corresponding depth image;
Step 2: improve the visual effect of the image that obtains through step 1, suppress garbage, and the image after improving is carried out Filtering Processing;
Step 3: represent the fruit respective peaks of diverse location to search in the gray scale frequency curve of the image that step 2 is obtained; And the variation characteristics of fruit center depth information in the combining image to the edge; Realization is successively cut apart the target area, obtains the one or more bianry image;
Step 4: the area ratio of the circularity of each connected region of one or more bianry image that obtains in the calculation procedure 3 and each connected region area and maximum area, through above-mentioned relative shape parameter is retrained, accomplish accurate identification to the fruit zone.
2. according to claim 1 a kind of based on fruit recognition methods on the tree of depth image; It is characterized in that; The method of the visual effect of the image that said improvement obtains through step 1 adopts the selectivity contrast enhancement process, because the gray-scale value that image histogram occurs belongs to background in a big way, directly ignores respective pixel; To contain between the setting gray area of important information and expand, accomplish tonal range by setting between gray area to the conversion full gray area.
3. according to claim 2ly a kind ofly it is characterized in that based on fruit recognition methods on the tree of depth image described Filtering Processing has adopted median filtering technology, the pixel window through 3 * 3 carries out Filtering Processing to image.
4. according to claim 2ly a kind ofly it is characterized in that based on fruit recognition methods on the tree of depth image the concrete steps of said step 3 are:
A) set gray threshold Gt, only consider that (x y)≤gray scale change of frequency in the Gt scope, promptly only discerns the top layer fruit to G ';
B) based on a) said scope, utilize the evaluation function of setting, through all peak dot positions that satisfy condition of evaluation function formula search, and reject the peak dot of its peak value less than maximal value 75%; If still show as the multimodal state,, further confirm the validity of peak dot through setting a peak separation threshold value Dt; As D≤Dt, explain that different fruit alternate position spikes distances are less, a back peak dot is confirmed as active position; Work as D>Dt, different fruit alternate position spikes are described apart from bigger, need two peak dots are all kept;
C) with peak dot position S
iAdd one and set side-play amount u as segmentation threshold; Be single peak dot if detect, and then between cut section [0, S
1+ u] comprise the fruit zone; If detect to a plurality of peak dots, can exist between a plurality of cut sections, [0, S
1+ u] expression ground floor fruit zone, [S
1+ (S
2-S
1)/2, S
2+ u] expression second layer fruit zone, and the like; Pixel value in the cut zone is given logical zero, and the pixel value outside the cut zone is given logical one; If the histogram of fruit image is unimodal situation, then obtain a width of cloth bianry image; If the histogram of fruit image is the multimodal situation, then obtain several binary map; I=1 wherein, 2 ..., n; N represents the sum of peak dot.
5. according to claim 4 a kind of based on fruit recognition methods on the tree of depth image; It is characterized in that the accurate recognition methods in said fruit zone is, after negating in the fruit zone that step 3 is obtained existing in the one or more bianry image; Applied mathematics morphology disposal route open operation; Promptly corrode after expansion earlier, the hole in the blank map is removed the burr among the figure; Utilize zone marker with different connected domains among the difference figure, calculate respectively each regional circularity R (j) and with the area of maximum area than A (j), wherein j representes the mark of connected domain; Through scope restriction, with A (j) to above characteristic parameter<1/>5A
MaxOr R (j)<0.6 area pixel assignment 0, other pixels remain unchanged, A
MaxRepresent the area maximal value of connected region; Realization is to the accurate extraction in fruit zone, and will obtain several as a result figure carry out exclusive disjunction, to obtain final complete fruit area image.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210258051.5A CN102831398B (en) | 2012-07-24 | 2012-07-24 | Tree apple recognition method based on depth image |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210258051.5A CN102831398B (en) | 2012-07-24 | 2012-07-24 | Tree apple recognition method based on depth image |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN102831398A true CN102831398A (en) | 2012-12-19 |
| CN102831398B CN102831398B (en) | 2014-12-10 |
Family
ID=47334526
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201210258051.5A Expired - Fee Related CN102831398B (en) | 2012-07-24 | 2012-07-24 | Tree apple recognition method based on depth image |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102831398B (en) |
Cited By (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103177257A (en) * | 2013-04-09 | 2013-06-26 | 江苏大学 | Image identification method and image classification method for coprinus comatus |
| CN103177445A (en) * | 2013-03-13 | 2013-06-26 | 浙江大学 | Outdoor tomato identification method based on subsection threshold image segmentation and light spot identification |
| CN103295018A (en) * | 2013-05-21 | 2013-09-11 | 常州大学 | Method for precisely recognizing fruits covered by branches and leaves |
| CN103310218A (en) * | 2013-05-21 | 2013-09-18 | 常州大学 | Precise recognition method of overlapped shielded fruits |
| CN103336946A (en) * | 2013-06-17 | 2013-10-02 | 浙江大学 | Binocular stereoscopic vision based clustered tomato identification method |
| CN103544493A (en) * | 2013-10-24 | 2014-01-29 | 江苏大学 | Fruit distinguishing and locating method with laser scanning and machine vision combined |
| CN104408407A (en) * | 2014-11-05 | 2015-03-11 | 江苏大学 | Night identification method of apple harvesting robot |
| CN104636716A (en) * | 2014-12-08 | 2015-05-20 | 宁波工程学院 | Method for identifying green fruits |
| CN104751117A (en) * | 2015-01-26 | 2015-07-01 | 江苏大学 | Lotus seedpod target image recognition method for picking robot |
| CN105835074A (en) * | 2015-01-12 | 2016-08-10 | 广明光电股份有限公司 | Teaching method for picking up and placing articles by robot arm |
| CN106951905A (en) * | 2017-02-16 | 2017-07-14 | 江苏大学 | Apple identification and localization method on a kind of tree based on TOF camera |
| CN107341491A (en) * | 2016-11-23 | 2017-11-10 | 华东交通大学 | A kind of navel orange recognition methods based on machine vision |
| CN108776793A (en) * | 2018-06-08 | 2018-11-09 | 江南大学 | A kind of crab male and female recognition methods based on region segmentation |
| CN108805893A (en) * | 2018-06-06 | 2018-11-13 | 甘肃农业大学 | apple image segmentation method and device |
| CN110751118A (en) * | 2019-10-25 | 2020-02-04 | 四川大学 | Algorithm for rapidly identifying multiple plant species |
| CN112326192A (en) * | 2020-10-28 | 2021-02-05 | 西南石油大学 | Quantitative characterization method for proppant velocity in proppant flat plate conveying experiment |
| CN113674840A (en) * | 2021-08-24 | 2021-11-19 | 平安国际智慧城市科技股份有限公司 | Medical image sharing method and device, electronic equipment and storage medium |
| CN114067206B (en) * | 2021-11-16 | 2024-04-26 | 哈尔滨理工大学 | Spherical fruit identification positioning method based on depth image |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109697444B (en) * | 2017-10-20 | 2021-04-13 | 中国移动通信有限公司研究院 | Object identification method and device based on depth image, equipment and storage medium |
-
2012
- 2012-07-24 CN CN201210258051.5A patent/CN102831398B/en not_active Expired - Fee Related
Non-Patent Citations (3)
| Title |
|---|
| 司永胜 等: "《基于机器视觉的苹果识别和形状特征提取》", 《农业机械学报》 * |
| 司永胜 等: "《苹果采摘机器人果实识别与定位方法》", 《农业机械学报》 * |
| 荀一 等: "《基于轮廓曲率的树上苹果自动识别》", 《江苏大学学报(自然科学版)》 * |
Cited By (28)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103177445A (en) * | 2013-03-13 | 2013-06-26 | 浙江大学 | Outdoor tomato identification method based on subsection threshold image segmentation and light spot identification |
| CN103177445B (en) * | 2013-03-13 | 2015-10-28 | 浙江大学 | Based on the outdoor tomato recognition methods of fragmentation threshold Iamge Segmentation and spot identification |
| CN103177257A (en) * | 2013-04-09 | 2013-06-26 | 江苏大学 | Image identification method and image classification method for coprinus comatus |
| CN103310218A (en) * | 2013-05-21 | 2013-09-18 | 常州大学 | Precise recognition method of overlapped shielded fruits |
| CN103295018A (en) * | 2013-05-21 | 2013-09-11 | 常州大学 | Method for precisely recognizing fruits covered by branches and leaves |
| CN103295018B (en) * | 2013-05-21 | 2016-04-13 | 常州大学 | A kind of branches and leaves block fruit precise recognition method |
| CN103310218B (en) * | 2013-05-21 | 2016-08-10 | 常州大学 | A kind of overlap blocks fruit precise recognition method |
| CN103336946A (en) * | 2013-06-17 | 2013-10-02 | 浙江大学 | Binocular stereoscopic vision based clustered tomato identification method |
| CN103336946B (en) * | 2013-06-17 | 2016-05-04 | 浙江大学 | A kind of cluster shape tomato recognition methods based on binocular stereo vision |
| CN103544493A (en) * | 2013-10-24 | 2014-01-29 | 江苏大学 | Fruit distinguishing and locating method with laser scanning and machine vision combined |
| CN104408407A (en) * | 2014-11-05 | 2015-03-11 | 江苏大学 | Night identification method of apple harvesting robot |
| CN104408407B (en) * | 2014-11-05 | 2019-04-02 | 江苏大学 | The means of night recognition of apple picking robot |
| CN104636716A (en) * | 2014-12-08 | 2015-05-20 | 宁波工程学院 | Method for identifying green fruits |
| CN104636716B (en) * | 2014-12-08 | 2018-04-13 | 宁波工程学院 | Green fruit recognition methods |
| CN105835074A (en) * | 2015-01-12 | 2016-08-10 | 广明光电股份有限公司 | Teaching method for picking up and placing articles by robot arm |
| CN105835074B (en) * | 2015-01-12 | 2020-07-10 | 达明机器人股份有限公司 | Teaching method for picking and placing object by robot arm |
| CN104751117A (en) * | 2015-01-26 | 2015-07-01 | 江苏大学 | Lotus seedpod target image recognition method for picking robot |
| CN104751117B (en) * | 2015-01-26 | 2018-04-24 | 江苏大学 | A kind of seedpod of the lotus target image identification method for picking robot |
| CN107341491A (en) * | 2016-11-23 | 2017-11-10 | 华东交通大学 | A kind of navel orange recognition methods based on machine vision |
| CN106951905A (en) * | 2017-02-16 | 2017-07-14 | 江苏大学 | Apple identification and localization method on a kind of tree based on TOF camera |
| CN108805893A (en) * | 2018-06-06 | 2018-11-13 | 甘肃农业大学 | apple image segmentation method and device |
| CN108776793A (en) * | 2018-06-08 | 2018-11-09 | 江南大学 | A kind of crab male and female recognition methods based on region segmentation |
| CN110751118A (en) * | 2019-10-25 | 2020-02-04 | 四川大学 | Algorithm for rapidly identifying multiple plant species |
| CN112326192A (en) * | 2020-10-28 | 2021-02-05 | 西南石油大学 | Quantitative characterization method for proppant velocity in proppant flat plate conveying experiment |
| CN112326192B (en) * | 2020-10-28 | 2021-07-20 | 西南石油大学 | Quantitative characterization method for proppant velocity in proppant flat plate conveying experiment |
| CN113674840A (en) * | 2021-08-24 | 2021-11-19 | 平安国际智慧城市科技股份有限公司 | Medical image sharing method and device, electronic equipment and storage medium |
| CN113674840B (en) * | 2021-08-24 | 2023-11-03 | 深圳平安智慧医健科技有限公司 | Medical image sharing method and device, electronic equipment and storage medium |
| CN114067206B (en) * | 2021-11-16 | 2024-04-26 | 哈尔滨理工大学 | Spherical fruit identification positioning method based on depth image |
Also Published As
| Publication number | Publication date |
|---|---|
| CN102831398B (en) | 2014-12-10 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102831398B (en) | Tree apple recognition method based on depth image | |
| CN105718945B (en) | Apple picking robot night image recognition method based on watershed and neural network | |
| CN1322465C (en) | Image segmentation and fingerprint line distance getting technique in automatic fingerprint identification method | |
| CN105975972A (en) | Bridge crack detection and characteristic extraction method based on image | |
| CN109255757B (en) | Method for segmenting fruit stem region of grape bunch naturally placed by machine vision | |
| CN107527343B (en) | A kind of agaricus bisporus stage division based on image procossing | |
| CN102676633A (en) | Method for automatically counting bacterial colonies | |
| CN105205821A (en) | Weld image segmentation method | |
| CN111798411B (en) | Concrete inner steel bar intelligent positioning method based on ground penetrating radar and deep learning | |
| CN102184550A (en) | Mobile platform ground movement object detection method | |
| CN103761743A (en) | Solid wood floor surface defect detecting method based on image fusion and division | |
| CN105809194A (en) | Method for translating SAR image into optical image | |
| CN106951905A (en) | Apple identification and localization method on a kind of tree based on TOF camera | |
| CN109840553A (en) | The extracting method and system, storage medium, electronic equipment for agrotype of ploughing | |
| CN104636722A (en) | Fast tracking recognition method for overlapped fruits by picking robot | |
| CN106157323A (en) | The insulator division and extracting method that a kind of dynamic division threshold value and block search combine | |
| CN110415241A (en) | A kind of surface of concrete structure quality determining method based on computer vision | |
| CA2578042A1 (en) | Candidate generation for lung nodule detection | |
| CN104951799A (en) | SAR remote-sensing image oil spilling detection and identification method | |
| CN102622598A (en) | SAR (Synthesized Aperture Radar) image target detection method based on zone markers and grey statistics | |
| CN102169580A (en) | Self-adaptive image processing method utilizing image statistic characteristics | |
| Tiede et al. | Process oriented object-based algorithms for single tree detection using laser scanning | |
| CN102750531A (en) | Method for detecting handwriting mark symbols for bill document positioning grids | |
| CN101976347A (en) | Method for recognizing overwater bridge in remote sensing image on basis of Mean Shift segmentation | |
| CN115661032A (en) | Intelligent pavement disease detection method suitable for complex background |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20141210 Termination date: 20180724 |