Ship identification method and system based on night low-light image, and storage medium

Abstract

The invention discloses a ship identification method and system based on a night low-light image, and a storage medium. The method comprises the following steps: selecting night low-light image data of a set solar zenith angle; performing median filtering on the original image in the scanning line direction, and subtracting the median filtering image from the original image to obtain a scanning direction peak index SSI; respectively counting SSI of each column in one day according to the lifting rails, arranging the SSI from small to large, and extracting SSIgt; 0, and the difference gt of adjacent radiance; the SSI of 0.2 is a background threshold value TV1; performing percentile regression on each column of TV1 to obtain a smooth background threshold TV2, and performing SSIgt; the pixels of the TV2 are marked as possible ship pixels; performing median filtering in the orbit direction, and calculating a peak index TSI, if SSIgt; if TSI is 1.5 * TSI, removing the corresponding pixels from the possible ship pixels; the ship identification method has the beneficial effects that ship identification can be realized according to the characteristics of MERSI / LLB night low-light images.

Description

Technical Field

[0001] The present invention belongs to the field of satellite remote sensing technology, and in particular relates to a ship identification method, system and storage medium based on nighttime low-light images. Background Art

[0002] Low-light-level remote sensing is crucial for observing and understanding the climate system. It can be used for weather cloud analysis, detecting the extent of sea ice and the turbidity of ocean waters, and detecting fires, volcanic ash plumes, and pyroclastic flows. Compared to other satellite data, low-light-level data is unique in its ability to directly reflect human activity. On land, low-light-level remote sensing can detect urban nighttime lights to study their physical and social attributes. At sea, low-light-level remote sensing can also detect specific vessels. International regulations for preventing collisions at sea require that all ships operating at night carry lights of a certain intensity. However, the required brightness is too weak for satellite detection. However, some commercial fishing vessels use high-intensity lights to exploit the phototaxis of fish and shrimp, making them visible through satellite low-light channels. As we all know, overfishing by humans can disrupt the ecological balance of the ocean. Traditional ship positioning systems such as the Automatic Identification System (AIS) or the Vessel Monitoring System (VMS) are usually only installed on large ships and can be manually turned off, which is not conducive to monitoring fishing activities on the high seas. In contrast, low-light remote sensing can identify ships in all sea areas and the results are more objective. Therefore, it is of great significance to the development of fisheries and the protection of the marine ecological environment.

[0003] Currently, the primary source of low-light-level data internationally is the Day / Night Band (DNB) channel of the Visible Infrared Imaging Radiometer Suite (VIIRS) onboard the US SNPP and NOAA-20 satellites. Based on VIIRS / DNB data, NOAA has developed a nighttime ship identification algorithm (Elvidge et al., 2015) and generates a Nighttime Boat Detection (NBD) product. The real-time NBD product is available for a fee. Because ships appear as bright pixels in low-light-level images, the key to ship identification lies in distinguishing these bright pixels from background noise.

[0004] In the existing technology, the VIIRS NBD algorithm can effectively reduce noise through a multi-detector set, examine the relationship between the peak index and the original radiance, and exclude whether the bright pixel is caused by natural light sources, and finally determine whether the bright pixel is a ship.

[0005] The FY-3E meteorological satellite launched by my country in 2021 is equipped with a Medium Resolution Imager (MERSI) and a Low Light Band (LLB). This enables Fengyun satellites to detect low light levels at night for the first time, making it the world's first operational meteorological satellite in a dawn-dusk orbit. Its descending orbit passes through local time at approximately 5:40 a.m. and ascending orbit passes at approximately 5:40 p.m., achieving full global data coverage every six hours in conjunction with the morning and afternoon satellite networks, meeting the six-hour assimilation window requirement for numerical weather forecasting. The FY-3E payload, equipped with a Medium Resolution Imager (MERSI) and a Low Light Band (LLB), has a subsatellite resolution of 1 km and a swath width of 2,500 km. The LLB has three gain stages: low, medium, and high, and can observe radiances ranging from 3.5×10-5 to 50 W / m2 / sr. Due to the differences between the MERSI / LLB and VIIRS / DNB detectors, MERSI / LLB does not have a detector set. MERSI / LLB cannot use filtering methods to make the noise on a scan line consistent like VIIRS / DNB does, making the VIIRS ship identification method unable to be directly applied to MERSI.

[0006] Therefore, there is an urgent need for a solution to realize ship recognition based on the characteristics of MERSI / LLB night-time low-light images. Summary of the Invention

[0007] In order to overcome the above-mentioned deficiencies in the prior art, the object of the present invention is to provide a method, system and storage medium for ship identification based on nighttime low-light images.

[0008] In a first aspect, the present invention provides a method for ship recognition based on nighttime low-light images, the method comprising:

[0009] Select nighttime low-light image data with a solar zenith angle greater than the set angle as nighttime observation data;

[0010] Perform median filtering on the original image in the scan line direction using a preset first pixel selection method, and obtain the scan direction peak index SSI by subtracting the median filtered image from the original image.

[0011] The SSI of each column of the day is counted in ascending and descending orbits and arranged in ascending order. The radiance of bright pixels will have a significant jump compared to background pixels. Then, the SSI with SSI>0 and the difference in radiance of adjacent SSI values>0.2nW / cm2 / sr is extracted as the initial background threshold TV1;

[0012] Perform percentile regression on each column TV1 to obtain the smooth background threshold TV2, and mark the pixels with SSI>TV2 as possible ship pixels;

[0013] Then, the preset second pixel selection method is used to perform median filtering on the original image in the track direction, and the peak index TSI in the track direction is calculated. If SSI>1.5×TSI, the corresponding pixel is eliminated from the possible ship pixel.

[0014] Preferably, the method further comprises:

[0015] Taking unit latitude as interval, calculate the standard deviation stdil of the observed radiance in each column and each latitude, and weight the TV2 by the ratio of stdil to the standard deviation stdi of the entire column as the final judgment threshold TV3;

[0016] Then, the pixels with SSI>TV3 are marked as possible ship pixels.

[0017] Preferably, the method further comprises:

[0018] Bright pixels caused by natural light sources are eliminated from the possible ship pixels.

[0019] Preferably, the preset first pixel selection method is 1×5 pixels, and the preset second pixel selection method is 5×1 pixels.

[0020] In a second aspect, the present invention further provides a ship identification system based on nighttime low-light images, the system comprising:

[0021] An acquisition module is used to select nighttime low-light image data with a solar zenith angle greater than a set angle as nighttime observation data;

[0022] A filtering module is used to perform median filtering on the original image in the scan line direction using a preset first pixel selection method, and obtain a scan direction peak index SSI by subtracting the median filtered image from the original image;

[0023] The processing module is used to count the SSI of each column of the day in ascending and descending orbits and arrange them in ascending order. The radiance of bright pixels will have a significant jump compared to background pixels. Then, the SSI with SSI>0 and the difference in radiance of adjacent SSI values>0.2nW / cm2 / sr is extracted as the initial background threshold TV1;

[0024] Identification module for:

[0025] Perform percentile regression on each column TV1 to obtain the smooth background threshold TV2, and mark the pixels with SSI>TV2 as possible ship pixels;

[0026] Then, the preset second pixel selection method is used to perform median filtering on the original image in the track direction, and the peak index TSI in the track direction is calculated. If SSI>1.5×TSI, the corresponding pixel is eliminated from the possible ship pixel.

[0027] In a third aspect, the present invention further provides a storage medium, wherein the computer storage medium stores a computer program, wherein the computer program includes program instructions, and when the program instructions are executed by a processor, the processor executes the method described in the first aspect.

[0028] The beneficial effects of the present invention are as follows: the present invention distinguishes background pixels from bright pixels by detecting radiance jumps, processes each column on the scan line, obtains corresponding thresholds to solve the problem of noise variation on the scan line; and calculates the peak index in the scan line direction and the track direction respectively to solve the inconsistency problem between scan lines; thereby realizing ship recognition based on the characteristics of MERSI / LLB night-time low-light images. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the following briefly describes the drawings required for the specific embodiments or the description of the prior art. Similar elements or parts are generally identified by similar reference numerals throughout the drawings. Elements or parts in the drawings are not necessarily drawn to scale.

[0030] Figure 1 A flowchart of a ship recognition method based on nighttime low-light images provided by an embodiment of the present invention;

[0031] Figure 2 A schematic diagram showing a significant jump in the radiance of a bright pixel relative to a background pixel provided by an embodiment of the present invention;

[0032] Figure 3 A schematic diagram of an initial background threshold TV1 and a smoothed background threshold TV2 provided in an embodiment of the present invention;

[0033] Figure 4 A distribution diagram of the standard deviation of ascending and descending orbit nighttime observations versus scanning angle and latitude provided by an embodiment of the present invention;

[0034] Figure 5 A local schematic diagram of L1 data provided by an embodiment of the present invention;

[0035] Figure 6 Based on Figure 5 Schematic diagram of the corresponding ship identification results;

[0036] Figure 7 This is a structural block diagram of a ship identification system based on nighttime low-light images provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0037] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0038] It will be understood that when used in this specification and the appended claims, the terms “comprises” and “comprising” indicate the presence of described features, integers, steps, operations, elements and / or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or groups thereof.

[0039] It should also be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the present invention. As used in the specification and appended claims, the singular forms "a," "an," and "the" are intended to include the plural forms unless the context clearly indicates otherwise.

[0040] Throughout this specification, references to "one embodiment," "an embodiment," "an example," or "an example" mean that a particular feature, structure, or characteristic described in connection with the embodiment or example is included in at least one embodiment of the present invention. Thus, the appearances of the phrases "in one embodiment," "in an embodiment," "an example," or "an example" in various places throughout this specification are not necessarily all referring to the same embodiment or example. Furthermore, the particular features, structures, or characteristics may be combined in any suitable combinations and / or subcombinations in one or more embodiments or examples.

[0041] Ascending: ascending track;

[0042] Descending: descending track;

[0043] Latitude: latitude;

[0044] Pixel: pixel;

[0045] index: The pixel serial number on the scan line.

[0046] It should be noted that, unless otherwise specified, the technical terms in this embodiment have the common meanings understood in the relevant technical field.

[0047] Reference Figure 1 This embodiment provides a method for identifying ships based on nighttime low-light images, the method comprising:

[0048] S101, selecting nighttime low-light image data with a solar zenith angle greater than a set angle as nighttime observation data; when applied, the set angle is 103°, and data with a solar zenith angle greater than 103° is selected as nighttime observation data;

[0049] S102, performing median filtering on the original image in the scan line direction using a preset first pixel selection method, and obtaining a scan direction peak index SSI by subtracting the median filtered image from the original image;

[0050] S103: Count the SSIs of each column of the day in ascending and descending orbits and arrange them in ascending order. The radiance of bright pixels will have a significant jump compared to background pixels. Then, the SSI with SSI>0 and the difference in radiance between adjacent SSIs>0.2nW / cm2 / sr is extracted as the initial background threshold TV1.

[0051] S104, perform percentile regression on each column TV1 to obtain the smooth background threshold TV2, and mark the pixels with SSI>TV2 as possible ship pixels;

[0052] S105 , a preset second pixel selection method is used to perform median filtering on the original image in the track direction, and the peak index TSI in the track direction is calculated. If SSI>1.5×TSI, the corresponding pixel is removed from the possible ship pixel.

[0053] In this embodiment, the preset first pixel selection method is 1×5 pixels, and the preset second pixel selection method is 5×1 pixels.

[0054] During implementation, for S102, since the difference between scanning lines caused by stray light is large, in order to obtain a stable background value, the original image is median filtered using 1×5 pixels, that is, only the scanning line direction is filtered, which can also partially eliminate the influence of stray light; the scanning direction peak index SSI is obtained by subtracting the median filtered image from the original image; it should be noted that the image is composed of its corresponding pixel values, which should be understood by those skilled in the art and will not be elaborated here.

[0055] The significant jump obtained in step S103 is shown in FIG. Figure 2 , background threshold TV1 see Figure 3 The figure shows the scattered points in the figure; the figure uses a column of pixels sorted from small to large SSI as an example.

[0056] Perform percentile regression (>95%) to obtain the smooth background threshold TV2, see Figure 3 The continuous curve obtained in the experiment is obtained; percentile regression is performed to obtain the characteristics of noise changes with the scanning column; different noise thresholds are set for different scanning angles of the instrument, that is, different columns of pixels in the image to extract bright pixels; other regression fitting algorithms can also be used during implementation.

[0057] Furthermore, considering that the noise has latitude variations, during implementation, the method further includes:

[0058] Taking unit latitude as interval, calculate the standard deviation stdil of the observed radiance in each column and each latitude, and weight the TV2 by the ratio of stdil to the standard deviation stdi of the entire column as the final judgment threshold TV3;

[0059] Then, the pixels with SSI>TV3 are marked as possible ship pixels.

[0060] Specifically, with 1 latitude as the interval, the standard deviation of the observed radiance in each column and each latitude is calculated. Figure 4 ), the final judgment threshold is TV3 = TV2 × stdil / stdi, which is the ratio of stdil to the standard deviation stdi of the entire column weighted by TV2.

[0061] When applied, statistical or machine learning methods can be used to fit the relationship between noise and latitude, and more complex weighting methods can be used to introduce the influence of latitude on noise.

[0062] In step S105, we take into account that since SSI only considers the scan line direction, bright lines along the track direction caused by stray light may also be considered bright pixels. Therefore, we further perform a median filter using 5×1 pixels to calculate the peak index (TSI) in the track direction. If SSI > 1.5×TSI, indicating that the pixel brightness in the track direction is not significant compared to nearby pixels, this pixel is removed from the list of possible ship pixels, and the remaining pixels are identified as ship pixels.

[0063] Furthermore, for natural light sources such as lightning and high-energy particles, based on the above solution, the method further includes:

[0064] Bright pixels caused by natural light sources are eliminated from the possible ship pixels; the remaining possible ship pixels are then identified as ship pixels; wherein, they can be identified based on their image features. For specific methods, refer to the existing technology: Fengyun-3E Low Light Observation and Nighttime Lights Product. IEEE Transactions on Geoscience and Remote Sensing, vol. 61, pp. 1-12, 4703612, doi: 10.1109 / TGRS.2023.3292236.

[0065] The above scheme was verified by using the L1 data local ( Figure 5 ) displays the ship identification results (i.e., Figure 6Visually, the recognition results all have corresponding bright pixels in the L1 data, and the accuracy rate is above 90%.

[0066] The above scheme distinguishes background pixels from bright pixels by detecting radiance jumps, processes each column on the scan line, and obtains corresponding thresholds to solve the problem of noise variation on the scan line. The peak index in the scan line direction and the track direction are calculated separately to solve the inconsistency problem between scan lines. This method can realize ship recognition based on the characteristics of MERSI / LLB night low-light imagery.

[0067] Reference Figure 7 Based on the same inventive concept, this embodiment further provides a ship identification system based on nighttime low-light images, the system comprising:

[0068] An acquisition module is used to select nighttime low-light image data with a solar zenith angle greater than a set angle as nighttime observation data;

[0069] A filtering module is used to perform median filtering on the original image in the scan line direction using a preset first pixel selection method, and obtain a scan direction peak index SSI by subtracting the median filtered image from the original image;

[0070] The processing module is used to count the SSI of each column of the day in ascending and descending orbits and arrange them in ascending order. The radiance of bright pixels will have a significant jump compared to background pixels. Then, the SSI with SSI>0 and the difference in radiance of adjacent SSI values>0.2nW / cm2 / sr is extracted as the initial background threshold TV1;

[0071] Identification module for:

[0072] Perform percentile regression on each column TV1 to obtain the smooth background threshold TV2, and mark the pixels with SSI>TV2 as possible ship pixels;

[0073] Then, the preset second pixel selection method is used to perform median filtering on the original image in the track direction, and the peak index TSI in the track direction is calculated. If SSI>1.5×TSI, the corresponding pixel is eliminated from the possible ship pixel.

[0074] In this embodiment, the identification module is further configured to:

[0075] Taking unit latitude as interval, calculate the standard deviation stdil of the observed radiance in each column and each latitude, and weight the TV2 by the ratio of stdil to the standard deviation stdi of the entire column as the final judgment threshold TV3;

[0076] Then, the pixels with SSI>TV3 are marked as possible ship pixels.

[0077] Furthermore, bright pixels caused by natural light sources are eliminated from the possible ship pixels.

[0078] It should be noted that for more specific workflows and beneficial effects of the identification system, please refer to the aforementioned method embodiment section. Therefore, its specific implementation method can refer to the description of the corresponding embodiments of each part, and will not be repeated here.

[0079] An embodiment of the present invention further provides a storage medium, wherein the computer storage medium stores a computer program, wherein the computer program includes program instructions, and when the program instructions are executed by a processor, the processor executes the method described in the method embodiment.

[0080] It should be understood that in the embodiments of the present invention, the processor is used to run or execute the operating system, various software programs, and its own instruction set stored in the internal memory. The processor may include, but is not limited to, one or more of a central processing unit (CPU), a general-purpose graphics processing unit (GPU), a microprocessor (MCU), a digital signal processor (DSP), a field programmable gate array (FPGA), and an application-specific integrated circuit (ASIC).

[0081] The computer-readable storage medium may include a cache, a high-speed random access memory (RAM), such as the common double data rate synchronous dynamic random access memory (DDR SDRAM), and may also include a non-volatile memory (NVRAM), such as one or more read-only memories (ROMs), disk storage devices, flash memory devices, or other non-volatile solid-state memory devices such as optical disks (CD-ROMs, DVD-ROMs), floppy disks or data tapes.

[0082] Those skilled in the art will appreciate that the modules and method steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the above description has generally described the composition and steps of each example according to their functions. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0083] In the several embodiments provided in this application, it should be understood that the described systems and methods can be implemented in other ways. For example, the system embodiments described above are merely illustrative. For example, the module division is merely a logical functional division. In actual implementation, other division methods may be used, such as combining or integrating multiple modules or components into another system, or ignoring or not implementing certain features.

[0084] In addition, the functional modules in various embodiments of the present invention may be integrated into a single processing unit, each module may exist physically separately, or two or more modules may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional modules.

[0085] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or make equivalent replacements for some or all of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present invention, and they should all be included in the scope of the claims and description of the present invention.

Claims

1. A ship recognition method based on nighttime low-light images, characterized in that: The method comprises: Select nighttime low-light image data with a solar zenith angle greater than the set angle as nighttime observation data; Perform median filtering on the original image in the scan line direction using a preset first pixel selection method, and obtain the scan direction peak index SSI by subtracting the median filtered image from the original image. The SSI of each column of the day is counted in ascending and descending orbits and arranged in ascending order. The radiance of bright pixels will have a significant jump compared to background pixels. Then, the SSI with SSI>0 and the difference in radiance of adjacent SSI values>0.2nW / cm2 / sr is extracted as the initial background threshold TV1; Perform percentile regression on each column TV1 to obtain the smooth background threshold TV2, and mark the pixels with SSI>TV2 as possible ship pixels; Then, the preset second pixel selection method is used to perform median filtering on the original image in the track direction, and the peak index TSI in the track direction is calculated. If SSI>1.5×TSI, the corresponding pixel is eliminated from the possible ship pixel.

2. The ship recognition method based on nighttime low-light images according to claim 1, characterized in that: The method further comprises: Taking unit latitude as interval, calculate the standard deviation stdil of the observed radiance in each column and each latitude, and weight the TV2 by the ratio of stdil to the standard deviation stdi of the entire column as the final judgment threshold TV3; Then, the pixels with SSI>TV3 are marked as possible ship pixels.

3. A ship recognition method based on nighttime low-light images according to claim 1 or 2, characterized in that: The method further comprises: Bright pixels caused by natural light sources are eliminated from the possible ship pixels.

4. The ship recognition method based on nighttime low-light images according to claim 1, characterized in that: The preset first pixel selection method is 1×5 pixels, and the preset second pixel selection method is 5×1 pixels.

5. A ship identification system based on nighttime low-light images, characterized in that: The system comprises: An acquisition module is used to select nighttime low-light image data with a solar zenith angle greater than a set angle as nighttime observation data; A filtering module is used to perform median filtering on the original image in the scan line direction using a preset first pixel selection method, and obtain a scan direction peak index SSI by subtracting the median filtered image from the original image; The processing module is used to count the SSI of each column of the day in ascending and descending orbits and arrange them in ascending order. The radiance of bright pixels will have a significant jump compared to background pixels. Then, the SSI with SSI>0 and the difference in radiance of adjacent SSI values>0.2nW / cm2 / sr is extracted as the initial background threshold TV1; Identification module for: Perform percentile regression on each column TV1 to obtain the smooth background threshold TV2, and mark the pixels with SSI>TV2 as possible ship pixels; Then, the preset second pixel selection method is used to perform median filtering on the original image in the track direction, and the peak index TSI in the track direction is calculated. If SSI>1.5×TSI, the corresponding pixel is eliminated from the possible ship pixel.

6. The ship identification system based on nighttime low-light images according to claim 5, characterized in that: The identification module is further used to: Taking unit latitude as interval, calculate the standard deviation stdil of the observed radiance in each column and each latitude, and weight the TV2 by the ratio of stdil to the standard deviation stdi of the entire column as the final judgment threshold TV3; Then, the pixels with SSI>TV3 are marked as possible ship pixels.

7. The ship identification system based on nighttime low-light images according to claim 6, characterized in that: The identification module is further used to: Bright pixels caused by natural light sources are eliminated from the possible ship pixels.

8. A storage medium storing a computer program, wherein the computer program includes program instructions, characterized in that: When the program instructions are executed by a processor, the processor is caused to perform the method according to any one of claims 1 to 4 .

Images