CN106845396B - Illegal fishing behavior identification method based on automatic image identification - Google Patents
Illegal fishing behavior identification method based on automatic image identification Download PDFInfo
- Publication number
- CN106845396B CN106845396B CN201710039306.1A CN201710039306A CN106845396B CN 106845396 B CN106845396 B CN 106845396B CN 201710039306 A CN201710039306 A CN 201710039306A CN 106845396 B CN106845396 B CN 106845396B
- Authority
- CN
- China
- Prior art keywords
- area
- pixel points
- image
- continuous point
- fishing
- 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.)
- Expired - Fee Related
Links
- 238000000034 method Methods 0.000 title claims abstract description 24
- 238000012544 monitoring process Methods 0.000 claims abstract description 18
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 claims description 23
- 238000012706 support-vector machine Methods 0.000 claims description 13
- 238000000605 extraction Methods 0.000 claims description 12
- 230000000007 visual effect Effects 0.000 claims description 8
- 238000004891 communication Methods 0.000 claims description 7
- 230000004927 fusion Effects 0.000 claims description 7
- 230000008030 elimination Effects 0.000 claims description 4
- 238000003379 elimination reaction Methods 0.000 claims description 4
- 238000013145 classification model Methods 0.000 claims description 3
- 239000000284 extract Substances 0.000 claims description 3
- 238000009434 installation Methods 0.000 claims description 3
- 230000006399 behavior Effects 0.000 abstract 3
- 241000251468 Actinopterygii Species 0.000 abstract 1
- 238000010586 diagram Methods 0.000 description 2
- 230000001105 regulatory effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
- G06V20/46—Extracting features or characteristics from the video content, e.g. video fingerprints, representative shots or key frames
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N7/00—Television systems
- H04N7/18—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
The invention discloses an illegal fishing behavior recognition method based on automatic image recognition, which is characterized in that images shot by a camera are processed in real time, whether illegal fishers exist in an area or not is judged, if the illegal fishers exist in the area, a voice module is started to play pre-recorded warning drive-away voice, the illegal fishers are reminded, fish schools are frightened to destroy the fishing behaviors of the illegal fishers, and meanwhile, shot monitoring pictures with the illegal fishing behaviors are sent to a manager. The invention can realize the automatic monitoring of the no-fishing area, reduce the expensive labor cost to the maximum extent and improve the management efficiency, has high monitoring accuracy and has great engineering application value.
Description
Technical Field
The invention relates to the field of image recognition and video monitoring, in particular to an illegal fishing behavior recognition method based on automatic image recognition.
Background
Fishing is an amateur well liked by modern people, and it is often seen that fishermen are fishing at the side of rivers and lakes, but fishing is not permitted in many places for regulatory needs and commercial interest, such as lakes in scenic spots, private contracted reservoirs.
The existing method for prohibiting the fisher from fishing in the illegal region mainly comprises the steps of arranging a warning board and arranging a manager to patrol, but the arrangement of the warning board in practice cannot effectively play a role in driving away the illegal fisher, and the arrangement of the manager to patrol wastes human resources greatly.
Disclosure of Invention
The invention aims to provide an illegal fishing behavior recognition method based on automatic image recognition, which can automatically recognize an illegal fisher in an image shot by a camera, send a voice warning to the fisher and send prompt information to a manager.
The technical scheme for realizing the purpose of the invention is as follows: an illegal fishing behavior identification method based on automatic image identification comprises the following steps:
step 7, detecting that the surrounding of the fishing rod is extracted by applying a background elimination algorithm in step 6, wherein the foreground target is different from the background recorded in the system;
step 8, counting the pixel point color distribution histogram of the foreground target obtained in the step 7, inputting the color distribution histogram into a trained support vector machine classifier, and judging whether the foreground target is a human body target;
and 9, if the fishing rod exists in the visual field in the step 6 and the human body target exists around the fishing rod in the step 8, performing voice early warning and sending the shot monitoring picture to the administrator, and if the fishing rod exists in the visual field in the step 6 and the human body target does not exist around the fishing rod in the step 8, only transmitting the shot monitoring picture to the administrator.
Compared with the prior art, the invention has the following remarkable advantages:
(1) the invention can realize the automatic monitoring of the fishing forbidden area, reduce the expensive labor cost to the maximum extent and improve the management efficiency; (2) the invention has high monitoring accuracy and great engineering application value.
Drawings
FIG. 1 is a functional block diagram of the system of the present invention.
Fig. 2 is a flow chart of the illegal fishing behavior recognition method based on automatic image recognition according to the present invention.
Fig. 3 is a schematic diagram of the steps of a trip-based connectivity algorithm.
Detailed Description
The illegal fishing behavior identification method based on automatic image identification adopts the technical scheme that: as shown in fig. 1, a camera, a processor module, a voice alarm device and a wireless communication module are arranged in an area where an illegal fisher needs to be monitored, an image shot by the camera is processed by the processor in real time, whether the illegal fisher exists in the area is judged, if the illegal fisher exists in the area, the voice module is started, pre-recorded information for warning the fisher is played, and the information of the illegal fisher is provided for an administrator in a wireless network mode by the communication module.
With reference to fig. 2, an illegal fishing behavior recognition method based on automatic image recognition includes the following steps:
step 7, detecting that the surrounding of the fishing rod is extracted by applying a background elimination algorithm in step 6, wherein the foreground target is different from the background recorded in the system;
step 8, counting the pixel point color distribution histogram of the foreground target obtained in the step 7, inputting the color distribution histogram into a trained support vector machine classifier, and judging whether the foreground target is a human body target;
and 9, if the fishing rod exists in the visual field in the step 6 and the human body target exists around the fishing rod in the step 8, performing voice early warning and sending the shot monitoring picture to the administrator, and if the fishing rod exists in the visual field in the step 6 and the human body target does not exist around the fishing rod in the step 8, only transmitting the shot monitoring picture to the administrator.
Further, the monitor image size in step 1 is compressed to the standard 640 × 320 image size by the image interpolation algorithm.
Further, step 2 specifically comprises:
adopting the kernels respectively as GxAnd GyThe edge extraction operator performs convolution on the image with the standard size obtained in the step (1), extracts pixel points of which the difference value with the color value of surrounding pixel points is larger than a set threshold value from the image, and regards the extracted pixel points as edge pixel points;
further, step 3 specifically comprises:
step 3-1, starting from the first line of the image, marking continuous adjacent edge Pixel points in each line as a continuous point set family, and simultaneously recording the starting point pixels of the continuous adjacent edge Pixel pointsstartAnd endpoint PixelendAnd the row number rowNum of the continuous point set, the serial number Fid of the continuous point set;
3-2, starting from the second line, if the continuous point set in the line and the continuous point set in the previous line have adjacent pixel points in the vertical direction, assigning the serial number of the continuous point set in the previous line to the continuous point set, if the continuous point set in the previous line and the continuous point set in the previous line are vertically adjacent, assigning the serial number of the continuous point set with smaller serial number in the previous line to the continuous point set, and writing the serial numbers of the continuous point sets in the previous line into an equivalent sequence equal (Fid)1，Fid2,..) indicating that these sets of consecutive points belong to the same connected region;
step 3-3, traversing all the continuous point sets, and replacing the serial numbers of the continuous point sets in all the equivalent sequences with the serial number of the minimum continuous point set in the equivalent sequences; and filling the serial numbers of the continuous point sets into the image, wherein one region with the same serial number in the image is a connected region.
Further, step 4 specifically includes:
the installation position of the camera is fixed, the relative positions of the water area and the shore in the image are fixed, and whether the pixel points in the communication area obtained in the step 3 are located in the water area or the shore is judged according to a preset value;
the length of the communicated region crossing the water area is the number of pixel row lines occupied by the pixels in the water area, the length of the communicated region crossing the shoreside is the number of pixel row lines occupied by the pixels on the shoreside, the total length is the sum of the lengths of the water area crossing and the shoreside crossing, and the average width is the average thickness of the pixels in the communicated region in the vertical direction.
Further, step 5 specifically comprises:
step 5-1, respectively endowing a trust value for the length of the communicated area across the water area, the length of the communicated area across the shore, the total length and the average width obtained in the step 4 according to the experimental result of the statistics of a large number of fishing rod samples, and forming 4 evidences;
step 5-2, m is respectively used for 4 trust values1，m2，m3，m4The evidence of (1) and the fusion rule of the evidence are expressed by formula (2):
wherein, A represents that the evidence is judged to be true,
A1,...Anrespectively representing that the evidences with the numbers of 1-n are true;
and 5-3, comparing the total trust value obtained in the step 5-2 with a set threshold value, and judging whether the connected area is a fishing rod.
Further, step 7 specifically comprises:
the background recorded in the system comprises background images shot at a fixed angle in different seasons and different time periods, and the background images recorded in the system are selected and called according to the current time;
the surrounding area of the fishing rod is the image area with the connected area 100 x 80 near the end pixel point on the bank.
Color value D under RGB color space of pixel points in 100 x 80 image areai，j(R, G, B) respectively corresponding to color values T of pixel points in the background picturei，j(R, G and B) are used as difference values, and the pixel points with the difference values larger than a set threshold value are extracted and used as foreground target pixel points Qi，j(R，G，B)。
Further, step 8 specifically comprises:
the format of the color distribution histogram is P { { r { { R { (R) }1，r2…r16}，{g1，g2，…g16}，{b1，b2，…b16And } respectively converting the color components of red, green and blue wave bands of the foreground target pixel point extracted in the step 7 into 16-level color scales, and counting the corresponding foreground target pixel in each scaleThe number of points, and the statistical result forms a color distribution histogram;
the method comprises the steps of classifying obtained foreground targets by adopting a support vector machine, training the support vector machine by adopting a real person demonstration, simulating the behavior of a fisherman by the real person, shooting the simulated behavior by a camera arranged at a specific position, counting the image area of the fisherman in an obtained video image sequence into a color distribution histogram form, inputting the color distribution histogram form into the support vector machine, and training to obtain a vector machine classification model.
In order to facilitate the specific use of the invention by the engineers using the patent of the invention, the following detailed description of the implementation steps of the invention is provided with reference to the drawings and the examples.
Examples
An illegal fishing behavior identification method based on automatic image identification comprises the following specific steps:
1.1 the camera adopted by the invention is a common monitoring camera or a high-definition camera for a smart phone, and a monitoring image of an RGB (Red, Green, Blue) color space is obtained from a real-time video recorded by the camera;
1.2 the monitor image size obtained in 1.1 is compressed to the standard 640 x 320 image size by an image interpolation algorithm.
2.1 the fishing rod edge extraction model referred to in the present invention means that the fishing rod has a slim characteristic and the fishing action is performed with the fishing rod placed obliquely to the ground, and the color of the fishing rod is clearly distinguished from the water surface and the bank.
2.2 the kernels are G respectivelyxAnd GyThe edge extraction operator performs convolution on the image with the standard size obtained in the step 1, extracts pixel points of which the difference value between the color values of the pixel points in the image and the color values of the surrounding pixel points is larger than a set threshold value, and regards the extracted pixel points as edgesAnd (6) pixel points.
And 3, classifying the edge pixel points extracted in the step 2 by using a stroke-based connectivity algorithm, and marking the edge pixel points which are connected up, down, left and right as a connected region, as shown in fig. 3, wherein the stroke-based connectivity algorithm is implemented by the following steps:
3.1 starting from the first line of the image, marking the continuous adjacent edge Pixel points in each line as a continuous point set family, and recording the starting point Pixel thereofstartAnd endpoint PixelendAnd the row number rowNum of the continuous point set, the serial number Fid of the continuous point set;
3.2 starting from the second line, if the set of consecutive points in the line and the set of consecutive points in the previous line have adjacent pixel points in the vertical direction, assigning to it the number of the set of consecutive points in the previous line, if they are vertically adjacent to the sets of consecutive points in the previous line, assigning to it the number of the set of consecutive points with smaller number in the previous line, and writing the numbers of the sets of consecutive points in the previous line into an equivalent sequence equal (Fid)1，Fid2,..) indicating that these sets of consecutive points belong to the same connected region;
3.3 traversing all the continuous point sets, and replacing the serial numbers of the continuous point sets in all the equivalent sequences with the serial number of the minimum continuous point set in the equivalent sequences; and filling the serial numbers of the continuous point sets into the image, wherein one region with the same serial number in the image is a connected region.
And 4, counting the parameters of the connected region obtained in the step 3, and calculating the length of the connected region across the water area, the length of the connected region across the shore, the total length and the average width.
4.1 the installation position of the camera in the invention is fixed, the relative positions of the water area and the shore in the image are fixed, and whether the pixel point in the communication area obtained in the step 3 is positioned in the water area or the shore is judged according to a preset value;
4.2 the length of the connected region crossing the water area is the number of pixel rows occupied by the pixels in the water area, the length of the connected region crossing the bank is the number of pixel rows occupied by the pixels on the bank, the total length is the sum of the lengths of the connected region crossing the water area and the connected region crossing the bank, and the average width is the average thickness of the pixels in the connected region in the vertical direction.
And 5, comprehensively judging whether the connected region is a fishing rod or not by adopting an evidence fusion theory on the parameters of the connected region obtained in the step 4.
And 5.1, respectively assigning a trust value to the length across the water area, the length across the shore, the total length and the average width of the connected area obtained in the step 4 according to the experimental result of counting a large number of fishing rod samples, and forming 4 evidences.
5.2 m for each of the 4 confidence values1，m2，m3，m4The evidence of (1), the fusion rule of evidence is expressed by equation (4):
wherein, A represents that the evidence is judged to be true,
A1,...Anrespectively representing that the evidences with the numbers of 1-n are true;
and 5.3, comparing the total trust value obtained in the step 5.2 with a set threshold value, and judging whether the connected area is a fishing rod.
And 6, repeating the steps 2 to 5 until the connected area is judged to be a fishing rod.
Step 7, applying a background elimination algorithm to extract foreground objects around the fishing rod detected in step 6 that are different with respect to the background recorded in the system.
7.1 the background recorded in the system comprises background images shot at a fixed angle in different seasons and different time periods, and the background images recorded in the system are selected and called according to the current time;
7.2 the area around the fishing rod referred to in the present invention is an image area of 100 x 80 near the end pixel point of the connected area on the bank;
7.3 color value D in RGB color space for pixels in 100X 80 image regioni，j(R, G, B) respectively corresponding to color values T of pixel points in the background picturei，j(R, G and B) are used as difference values, and the pixel points with the difference values larger than a set threshold value are extracted and used as foreground target pixel points Qi，j(R，G，B)。
And 8, counting the pixel point color distribution histogram of the foreground target obtained in the step 7, inputting the color distribution histogram into a trained support vector machine classifier, and judging whether the foreground target is a human body target.
8.1 color distribution histogram of the format P { { r1，r2…r16}，{g1，g2，…g16}，{b1，b2，…b16Converting the color components of red, green and blue wave bands of the foreground target pixel points extracted in the step 7 into 16-level color scales respectively, counting the number of corresponding foreground target pixel points in each scale, wherein the counted result forms a color distribution histogram in the invention;
8.2 the invention adopts a support vector machine to classify the obtained foreground targets, the process of training the support vector machine is completed by adopting real person demonstration, the behavior of a fisherman is simulated by the real person, a camera arranged at a specific position shoots the simulated behavior, the image area of the fisherman in the obtained video image sequence is counted into a color distribution histogram form and is input into the support vector machine, and a vector machine classification model is obtained by training.
And step 9: if step 6 judges that there is a fishing rod in the visual field and step 8 judges that there is a human target around the fishing rod, the voice module sends out a driving-away warning sound and sends the photographed monitoring picture to the administrator through the wireless communication module, and if step 6 judges that there is a fishing rod in the visual field and step 8 judges that there is no human target around the fishing rod, only the photographed monitoring picture is transmitted to the administrator through the wireless communication module and the administrator judges whether there is a need to send out a voice driving-away sound.
The illegal fishing behavior identification method based on automatic image identification identifies the picture shot by the camera, and once an illegal fisher is found, the alarm driving-off sound is sent out, the sent alarm driving-off sound can warn the illegal fisher and can also forcibly destroy the fishing behavior of the illegal fisher, and meanwhile, the system can send monitoring information to a manager.
Claims (8)
1. An illegal fishing behavior identification method based on automatic image identification is characterized by comprising the following steps:
step 1, extracting a video image sequence from a camera installed in a no-fishing area in real time;
step 2, establishing an edge extraction model of the fishing rod, and extracting image edge pixel points from the input video image sequence by using an edge extraction operator;
step 3, classifying the edge pixel points extracted in the step 2 by using a stroke-based connectivity algorithm, and marking the edge pixel points which are connected up, down, left and right as a connected region;
step 4, calculating the length of the connected area across the water area, the length of the connected area across the shore, the total length and the average width;
step 5, comprehensively judging whether the connected region is a fishing rod or not by adopting an evidence fusion theory for the parameters of the connected region obtained in the step 4;
step 6, repeating the steps 2 to 5 until the connected area is judged as a fishing rod;
step 7, detecting that the surrounding of the fishing rod is extracted by applying a background elimination algorithm in step 6, wherein the foreground target is different from the background recorded in the system;
step 8, counting the pixel point color distribution histogram of the foreground target obtained in the step 7, inputting the color distribution histogram into a trained support vector machine classifier, and judging whether the foreground target is a human body target;
and 9, if the fishing rod exists in the visual field in the step 6 and the human body target exists around the fishing rod in the step 8, performing voice early warning and sending the shot monitoring picture to the administrator, and if the fishing rod exists in the visual field in the step 6 and the human body target does not exist around the fishing rod in the step 8, only transmitting the shot monitoring picture to the administrator.
2. An illegal fishing behavior recognition method based on automatic image recognition according to claim 1, characterized in that the size of the monitoring image in step 1 is compressed to a standard 640 x 320 image size by an image interpolation algorithm.
3. The illegal fishing behavior recognition method based on automatic image recognition according to claim 1, wherein the step 2 is specifically:
adopting the kernels respectively as GxAnd GyThe edge extraction operator performs convolution on the image with the standard size obtained in the step (1), extracts pixel points of which the difference value with the color value of surrounding pixel points is larger than a set threshold value from the image, and regards the extracted pixel points as edge pixel points;
4. an illegal fishing behavior recognition method based on automatic image recognition according to claim 1, characterized in that step 3 is specifically:
step 3-1, starting from the first line of the image, marking continuous adjacent edge Pixel points in each line as a continuous point set family, and simultaneously recording the starting point pixels of the continuous adjacent edge Pixel pointsstartAnd endpoint PixelendAnd the row number rowNum of the continuous point set, the serial number Fid of the continuous point set;
3-2, if the continuous point set in the line and the continuous point set in the previous line have adjacent pixel points in the vertical direction from the second line, assigning the serial number of the continuous point set in the previous line to the continuous point set, if the continuous point set is vertically adjacent to the plurality of continuous point sets in the previous line, assigning the serial number of the continuous point set with smaller serial number in the previous line to the continuous point set, and assigning the serial number of the continuous point set with smaller serial number in the previous line to the continuous point set in the previous lineIs written into an equivalent sequence equivalent (Ftd) of the number of the plurality of sets of consecutive points of1，Ftd2… …), indicating that these sets of consecutive points belong to the same connected region;
step 3-3, traversing all the continuous point sets, and replacing the serial numbers of the continuous point sets in all the equivalent sequences with the serial number of the minimum continuous point set in the equivalent sequences; and filling the serial numbers of the continuous point sets into the image, wherein one region with the same serial number in the image is a connected region.
5. The illegal fishing behavior recognition method based on automatic image recognition according to claim 1, wherein the step 4 is specifically:
the installation position of the camera is fixed, the relative positions of the water area and the shore in the image are fixed, and whether the pixel points in the communication area obtained in the step 3 are located in the water area or the shore is judged according to a preset value;
the length of the communicated region crossing the water area is the number of pixel row lines occupied by the pixels in the water area, the length of the communicated region crossing the shoreside is the number of pixel row lines occupied by the pixels on the shoreside, the total length is the sum of the lengths of the water area crossing and the shoreside crossing, and the average width is the average thickness of the pixels in the communicated region in the vertical direction.
6. The illegal fishing behavior recognition method based on automatic image recognition according to claim 1, wherein the step 5 is specifically:
step 5-1, respectively endowing a trust value for the length of the communicated area across the water area, the length of the communicated area across the shore, the total length and the average width obtained in the step 4 according to the experimental result of the statistics of a large number of fishing rod samples, and forming 4 evidences;
step 5-2, m is respectively used for 4 trust values1，m2，m3，m4The evidence of (1) and the fusion rule of the evidence are expressed by formula (2):
wherein, A represents that the evidence is judged to be true,
A1,...Anrespectively representing that the evidences with the numbers of 1-n are true;
and 5-3, comparing the total trust value obtained in the step 5-2 with a set threshold value, and judging whether the connected area is a fishing rod.
7. The illegal fishing behavior recognition method based on automatic image recognition according to claim 1, wherein the step 7 is specifically:
the background recorded in the system comprises background images shot at a fixed angle in different seasons and different time periods, and the background images recorded in the system are selected and called according to the current time;
the surrounding area of the fishing rod is an image area of 100 x 80 near the end pixel point of the communicated area on the bank;
color value D under RGB color space of pixel points in 100 x 80 image areai，j(R, G, B) respectively corresponding to color values T of pixel points in the background picturet，f(R, G and B) are used as difference values, and the pixel points with the difference values larger than a set threshold value are extracted and used as foreground target pixel points Qt，f(R，G，B)。
8. The illegal fishing behavior recognition method based on automatic image recognition according to claim 1, wherein the step 8 is specifically:
the format of the color distribution histogram is P { { r { { R { (R) }1，r2…r16}，{g1，g2，…g16}，{b1，b2，…b16Converting the color components of red, green and blue wave bands of the foreground target pixel points extracted in the step 7 into 16-level color scales respectively, counting the number of corresponding foreground target pixel points in each scale, and forming a color distribution histogram by the counted result;
the method comprises the steps of classifying obtained foreground targets by adopting a support vector machine, training the support vector machine by adopting a real person demonstration, simulating the behavior of a fisherman by the real person, shooting the simulated behavior by a camera arranged at a specific position, counting the image area of the fisherman in an obtained video image sequence into a color distribution histogram form, inputting the color distribution histogram form into the support vector machine, and training to obtain a vector machine classification model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710039306.1A CN106845396B (en) | 2017-01-18 | 2017-01-18 | Illegal fishing behavior identification method based on automatic image identification |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710039306.1A CN106845396B (en) | 2017-01-18 | 2017-01-18 | Illegal fishing behavior identification method based on automatic image identification |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN106845396A CN106845396A (en) | 2017-06-13 |
| CN106845396B true CN106845396B (en) | 2020-05-22 |
Family
ID=59124446
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710039306.1A Expired - Fee Related CN106845396B (en) | 2017-01-18 | 2017-01-18 | Illegal fishing behavior identification method based on automatic image identification |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106845396B (en) |
Families Citing this family (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108520511A (en) * | 2018-03-19 | 2018-09-11 | 中国海洋大学 | A kind of underwater fish target detection and identification method based on fish finder |
| CN108648211B (en) * | 2018-05-16 | 2020-05-01 | 思百达物联网科技(北京)有限公司 | Small target detection method, device, equipment and medium based on deep learning |
| CN109040693B (en) * | 2018-08-31 | 2020-11-10 | 上海赛特斯信息科技股份有限公司 | Intelligent alarm system and method |
| CN111160255B (en) * | 2019-12-30 | 2022-07-29 | 成都数之联科技股份有限公司 | Fishing behavior identification method and system based on three-dimensional convolution network |
| CN111583265A (en) * | 2020-05-07 | 2020-08-25 | 赛特斯信息科技股份有限公司 | Method for realizing phishing behavior detection processing based on codec structure and corresponding semantic segmentation network system |
| CN111695492A (en) * | 2020-06-10 | 2020-09-22 | 国网山东省电力公司电力科学研究院 | Method and system for detecting fishing hidden danger of power transmission line |
| CN112233353A (en) * | 2020-09-24 | 2021-01-15 | 国网浙江兰溪市供电有限公司 | Artificial intelligence-based anti-fishing monitoring system and monitoring method thereof |
| CN114596536A (en) * | 2022-05-07 | 2022-06-07 | 陕西欧卡电子智能科技有限公司 | Unmanned ship coastal inspection method and device, computer equipment and storage medium |
| CN115331386B (en) * | 2022-10-13 | 2022-12-27 | 合肥中科类脑智能技术有限公司 | Prevent fishing detection alarm system based on computer vision |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101090482A (en) * | 2006-06-13 | 2007-12-19 | 唐琎 | Driver fatigue monitoring system and method based on image process and information mixing technology |
| CN104637059A (en) * | 2015-02-09 | 2015-05-20 | 吉林大学 | Night preceding vehicle detection method based on millimeter-wave radar and machine vision |
| CN105975913A (en) * | 2016-04-28 | 2016-09-28 | 武汉大学 | Road network extraction method based on adaptive cluster learning |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20060052923A1 (en) * | 2004-09-03 | 2006-03-09 | Eaton Corporation (Rj) | Classification system and method using relative orientations of a vehicle occupant |
| JP4450799B2 (en) * | 2006-03-10 | 2010-04-14 | 富士フイルム株式会社 | Method for controlling target image detection apparatus |
-
2017
- 2017-01-18 CN CN201710039306.1A patent/CN106845396B/en not_active Expired - Fee Related
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101090482A (en) * | 2006-06-13 | 2007-12-19 | 唐琎 | Driver fatigue monitoring system and method based on image process and information mixing technology |
| CN104637059A (en) * | 2015-02-09 | 2015-05-20 | 吉林大学 | Night preceding vehicle detection method based on millimeter-wave radar and machine vision |
| CN105975913A (en) * | 2016-04-28 | 2016-09-28 | 武汉大学 | Road network extraction method based on adaptive cluster learning |
Non-Patent Citations (2)
| Title |
|---|
| "基于D-S证据理论的信息融合图像识别";张逵 等;《上海海事大学学报》;20120930;第33卷(第3期);81-86 * |
| "基于神经网络和证据理论的图像目标识别研究";黄金 等;《计算机仿真》;20051130;第22卷(第11期);184-186 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN106845396A (en) | 2017-06-13 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106845396B (en) | Illegal fishing behavior identification method based on automatic image identification | |
| CN107105310B (en) | Human image replacing method and device in video live broadcast and recording and broadcasting system | |
| Verdoliva | Extracting camera-based fingerprints for video forensics | |
| CN105989367B (en) | Target Acquisition method and apparatus | |
| CN111898581B (en) | Animal detection method, apparatus, electronic device, and readable storage medium | |
| KR101204259B1 (en) | A method for detecting fire or smoke | |
| CN109981943A (en) | Picture pick-up device, image processing equipment, control method and storage medium | |
| CN112750104B (en) | Method and device for automatically matching optimal camera by monitoring ship through multiple cameras | |
| CN110781853A (en) | Crowd abnormality detection method and related device | |
| CN110288623B (en) | Data compression method for unmanned aerial vehicle maritime net cage culture inspection image | |
| US20170068855A1 (en) | Method and system for facilitating real time detection of linear infrastructural objects by aerial imagery | |
| CN111008563A (en) | Seed germination detection method and device in dark scene and readable storage medium | |
| CN116109954B (en) | House potential safety hazard identification method, device and storage medium | |
| CN112163544A (en) | Method and system for judging random placement of non-motor vehicles | |
| WO2020168515A1 (en) | Image processing method and apparatus, image capture processing system, and carrier | |
| CN110942577A (en) | Machine vision-based river sand stealing monitoring system and method | |
| CN108090426A (en) | A kind of group rearing pig personal identification method based on machine vision | |
| CN112528823B (en) | Method and system for analyzing batcharybus movement behavior based on key frame detection and semantic component segmentation | |
| US6526169B1 (en) | Histogram-based segmentation of objects from a video signal via color moments | |
| CN104899854B (en) | The detection method and device of heap grain altitude line | |
| CN110852172B (en) | Method for expanding crowd counting data set based on Cycle Gan picture collage and enhancement | |
| CN116343100B (en) | Target identification method and system based on self-supervision learning | |
| CN112233353A (en) | Artificial intelligence-based anti-fishing monitoring system and monitoring method thereof | |
| Jin et al. | Fusing Canny operator with vibe algorithm for target detection | |
| JP2001167273A5 (en) |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20200522 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |