Separation of leaves and woody materials is crucial for estimation of biophysical attributes of trees such as leaf area index and above-ground biomass. Segmentation based on images from traditional cameras is difficult in low-light level conditions like high canopy density areas and understory vegetation in rain forests. Point clouds from terrestrial laser scanning (TLS) LIDAR are also used for canopy quantitative analysis, but it suffers from low spatial-resolution for leafwood separation. To solve the problems mentioned above, we present a method of wood-leaf separation method based on a dual-wavelength active range-gated imaging system to separate leaves, woody elements, and background. In our method, two overlapped near infrared gated images at the wavelength of 808nm are obtained with background filtered by gated viewing, and a green-channel image is grasped at the illumination of 530nm LED. Then through data preprocessing, these images are input into our separation algorithm. Our separation method uses a peak-search formula to find the target peak in the histogram of an image, and the threshold is the local minimum at the right of the target peak. After segmentation by thresholding, a woody area mask is obtained. Combined with the point clouds reconstructed from gated images, separation on the point clouds is available. We have collected images of vegetation and performed manual separation to test our method. The results show that our method is capable to make accurate classification of leaves, woody elements and background.
Underwater in-situ darkfield microscopy imaging can grasp excellent resolution and high contrast images for observing transparent and living plankton. However, its limited field of view and shallow depth of field leads to a large amount of redundant data during image acquisition. Manual extraction of regions of interest (ROIs) imposes significant labor and time costs. Therefore, we develop an automatic ROI extraction method to obtain high-quality in-focus individual plankton images from raw images. A multiscale underwater in-situ microscopy system (MUIMS)based on darkfield imaging has been established. The system allows for the monitoring of plankton in the size range from 10μm to 1cm. A series of preprocessing steps are performed on the raw darkfield images to make edge detection and segmentation of plankton regions. This can realize rapid extraction of ROIs from the raw images. A dataset is constructed from these ROIs, and each image is labeled to indicate whether it is a high-quality image. By using a transfer learning strategy, we improve a pre-trained ResNet-18 model and train it to accurately classify high-quality ROIs. The experimental results show that this method can rapidly extract useful information from raw images, achieving an accuracy of over 90% in judging the quality of ROIs. This research is beneficial for building a dataset of high-quality plankton, which provides a crucial data foundation for ecological research and marine biodiversity conservation.
An underwater microbubbles detection and identification method based on dark field imaging is presented, focusing on measurement of the characteristics of microbubbles with high-resolution image capture. The target microbubbles in sampling volume of interest are illuminated by a laser sheet, which is matched with the depth of field. Images are captured in-focus under an observation angle of 90 deg to implement dark field detection, which suppresses the influence of interferential light, such as backlighting, scattering light, and blur caused by defocus. In this configuration, we propose to measure bubble size based on the distance of peak values of two glare points formed by the diffraction and reflection lights. The modified Resnet18 is employed to recognize the bubble images, and the number of bubbles with different diameters in bubble images is counted by template matching. Experimental results show that this dark field configuration allows high resolution imaging and evaluation of the diameter scale of the microbubbles. The precision of bubble recognition and quantification of bubbles within a certain range achieves more than 90%.
Underwater three-dimensional gated range-intensity correlation imaging (3D GRICI) can obtain two-dimensional images with large target-to-background contrast by suppressing backscatter and background noise outside volume of interest (VOI), and simultaneously reconstruct 3D images by the range-intensity correlation algorithm. However, it is still affected by the sub-backscatter noise from VOI in turbid water, resulting in short 3D detection distance and low contrast images. Therefore, in this paper the optical polarization is used for 3D GRICI. Due to the polarization-preserving property of water backscattering, the sub-backscatter noise can be removed from gate images by polarization. Experimental results show that when the water attenuation coefficient is less than or equal to the critical attenuation coefficient of c0, the polarization has no effect on improving detection distance and imaging quality of 3D GRICI; when the water attenuation coefficient is greater than c0, the polarization is helpful for improving performance of 3D GRICI. This research is conducive to optimize the applications of 3D GRICI in turbid water.
It needs a triggered time to open or close optical gate of the ICCD. The duration time is defined as irising effect time. Although it only lasts a few nanoseconds and often be ignored by users, it can still interfere with the results in some applications such as fluorescence lifetime imaging and gated imaging. This paper proposes a fitting algorithm to correct the irising effect of the image intensifier. This method obtains fitting matrices of different gates through a series delay images of ICCD. Then through these fitting matrices, the imaging pictures are effectively corrected. The advantages of this method are low cost, high efficiency, and simplicity. The verification experiment of this paper is to write letters with a highlighter. The correction algorithm can clearly restore time-resolve image, and significantly improve the contrast of the image.
There are a large number of bubbles in ocean, such as H2S, CH4 and CO2 in hydrothermal and cold springs. Bubbles in
seawater are moving targets with small scale and high transparency, and it is difficult to obtain their clear optical images.
In order to better image tiny bubbles, it is necessary to explore the optical characteristics of bubbles under different angles
of incident light. By building a multi-angle optical characteristic analysis device for underwater bubbles, this paper collects
bubble images at different incident angles, and analyzes the image characteristics of tiny bubbles. The typical bubble
patterns are "hamburger" and "doughnut". In experiment, the bubble groups at different angles were collected and analyzed,
and an orthogonal narrow slice imaging scheme was proposed. The optical slice imaging can obtain clear optical images
of bubbles, which is conducive to quantitative analysis of bubble identification and counting in the later stage.
Gated range-intensity correlation 3D imaging is a fast high-resolution 3D imaging technology based on range-gated imaging, and has great potential in underwater imaging, marine life detection and topographic mapping. Pixel gray information between different frame-type 2D images is used to obtain 3D image and distance information. However, for moving platforms (such as underwater vehicles), and moving targets (such as zooplankton, fish, etc.), relative motion will cause two frames of images to be misaligned, eventually leading to failure of 3D inversion. To solve the problem, the inter-frame feature matching for underwater 3D gated range-intensity correlation imaging method is proposed. Feature pairs are matched by the bi-directional feature matching, and mean value of the feature-pair coordinate differences is optimized, the method can solve feature point-to-coordinate deviation caused by water body noise, and improve the robustness of feature point matching. Experimental verification on pools and lakes shows that the proposed method is effective and can realize three-dimensional imaging of relative moving targets.
Underwater high-resolution optical imaging is widely used in marine resource development, ecological monitoring, underwater archaeology, underwater search and rescue, and marine scientific research. However, small field of view (FOV) in optical imaging is ineffective for large scene mapping. Underwater image stitching aims to generate a highresolution panoramic image containing more information from a series of small-FOV and high-resolution images. This paper proposes an underwater optical image stitching method, including image alignment algorithm, optimal seam algorithm, Hue-Saturation-Value correction algorithm based on histogram matching, and multi-resolution fusion algorithm. The experimental results show that the proposed algorithms are effective for targets in different environments, and the panorama can be produced without artifacts or visible seams.
High-resolution 3D optical imaging is important for unmanned underwater vehicles (UUVs) in the applications of target detection and recognition, underwater engineering, automatic navigation, scientific research, and natural resources exploration. Compared with stereo camera and 3D laser scanning imaging, 3D range-gated imaging (3D RGI) has longer detection range and higher spatial resolution at the same time. This paper presents a survey on 3D range-gated imaging methods for underwater detection. Up to now, there are two main methods developed, including time slicing method and gated range intensity correlation method. The literature about 3D RGI has been reviewed in different detection applications for UUV. We also introduce our works in 3D gated range-intensity correlation imaging for fishing net detection and marine life in-situ detection. This paper is beneficial for the 3D RGI technique in underwater detection applications for UUV.
Aiming at the practical application requirements of small dark target recognition for underwater unmanned aerial vehicles, a underwater laser gating imaging target recognition network based on convolutional neural network is designed to classify and identify underwater multiple targets. The integrated tool HLS transplants the network into the FPGA for circuit implementation. Firstly, the algorithm is designed to verify the realization of the convolutional neural network. Then the underwater target recognition experiment is carried out on the implemented convolutional neural network circuit. The network identification accuracy rate is 94% for the three types of underwater target used in the experiment, which verifies the feasibility of convolutional neural network implementation in FPGA.
KEYWORDS: 3D image processing, 3D acquisition, Target detection, Near infrared, Night vision, Stereoscopy, Pulsed laser operation, Gated imaging, Imaging systems, Night vision systems
Traditional NIR laser night vision systems can only obtain 2D images without target range information, and are also easily affected by fog, rain, snow and foreground/background. To solve the problems above, 3D laser night vision based on range-gated imaging has been developed. This paper reviews 3D range-gated imaging advances and focuses on 3D rangeintensity correlation imaging (GRICI) due to its better real time performance and higher spatial resolution. In GRICI systems, the typical illuminator is eye-invisible pulsed semiconductor laser, and the image sensor chooses gated ICCD or ICMOS with mega pixels and ns-scaled gate time. To realize 3D night vision, two overlapped gate images with trapezoidal or triangular range-intensity profiles are grasped by synchronizing the puled laser and the gated sensor. The collapsed range is reconstructed by the range-intensity correlation algorithm, and furthermore 2D and 3D images can both be obtained at the same frame rate. We have established 3D NIR night vision systems based on triangular GRICI, and the experimental results demonstrate that 3D images realize target extraction from background and through windows or smoke. The range resolution minimum is about less than 0.2m at the range of 1km in our GRICI-NV3000, and the range maximum of 3D imaging is about 5km in our GRICI-NV6000.
The underwater optical vision guidance technique is important for AUV short-range docking. Most of the docking lights are blue or green LED lights. However, LED lights have large divergence angle and cannot meet long range optical guidance. In this paper, we designed a blue laser diode docking light named as Beijixing with adjustable divergence angle and carried out the experiment. The divergence angle variation model of the docking light propagating in water is established according to the optical properties of water and the beam-spread function. The experimental results show that at a distance of 18m, when the divergence angle is 0.06° with strong directivity, the images captured by Nano SeaCam appear saturated, and the spot images captured by StarfishCam are light columns, which are difficulty for spot centroid extraction. When the angle is 10°, the light spots captured by Nano-SeaCam overlap, but StarfishCam can distinguish the spots. The research is beneficial for the design of docking lights for underwater optical vision guidance in AUV docking.
Camera traps are commonly used in wildlife monitoring. Traditionally camera traps only capture 2D images of wildlife moving in front of them. However, size information of wildlife is lost, which is vital to determine their ages and genders. To solve this problem, this paper develops a binocular camera trap based on stereo imaging for wildlife detection. The camera trap consists of two cameras, motion sensors, a photosensitive sensor and infrared illumination with the central wavelength of 940nm. Motion sensors output triggers to cameras when animals move past, and then pictures are captured from two different perspectives simultaneously. Meanwhile the photosensitive sensor perceives ambient illumination to control infrared illumination. In this way, the camera trap provides both 2D images of wildlife and their size information obtained by binocular vision. In addition, different from normal binocular cameras placed horizontally, these two cameras are set vertically for the convenience of installation and the expansion of dynamic measure range. As verification, we develop a prototype binocular camera trap to measure a human’s height that is 178cm, and the estimation error approaches 2cm at the distance of 5m.
KEYWORDS: 3D image processing, 3D metrology, Pulsed laser operation, Super resolution, Cameras, 3D acquisition, Imaging systems, Reconstruction algorithms, 3D modeling, Image processing
In this paper, we proposed a method of canopy reconstruction and measurement based on 3D super resolution range-gated imaging. In this method, high resolution 2D intensity images are grasped by active gate imaging, and 3D images of canopy are reconstructed by triangular-range-intensity correlation algorithm at the same time. A range-gated laser imaging system(RGLIS) is established based on 808 nm diode laser and gated intensified charge-coupled device (ICCD) camera with 1392´1040 pixels. The proof experiments have been performed for potted plants located 75m away and trees located 165m away. The experiments show it that can acquire more than 1 million points per frame, and 3D imaging has the spatial resolution about 0.3mm at the distance of 75m and the distance accuracy about 10 cm. This research is beneficial for high speed acquisition of canopy structure and non-destructive canopy measurement.
Moving target detection is important for the application of target tracking and remote surveillance in active range-gated laser imaging. This technique has two operation modes based on the difference of the number of pulses per frame: stroboscopic mode with the accumulation of multiple laser pulses per frame and flash mode with a single shot of laser pulse per frame. In this paper, we have established a range-gated laser imaging system. In the system, two types of lasers with different frequency were chosen for the two modes. Electric fan and horizontal sliding track were selected as the moving targets to compare the moving blurring between two modes. Consequently, the system working in flash mode shows more excellent performance in motion blurring against stroboscopic mode. Furthermore, based on experiments and theoretical analysis, we presented the higher signal-to-noise ratio of image acquired by stroboscopic mode than flash mode in indoor and underwater environment.
3D range-gated superresolution imaging is a novel 3D reconstruction technique for target detection and recognition with good real-time performance. However, for moving targets or platforms such as airborne, shipborne, remote operated vehicle and autonomous vehicle, 3D reconstruction has a large error or failure. In order to overcome this drawback, we propose a method of stereo matching for 3D range-gated superresolution reconstruction algorithm. In experiment, the target is a doll of Mario with a height of 38cm at the location of 34m, and we obtain two successive frame images of the Mario. To confirm our method is effective, we transform the original images with translation, rotation, scale and perspective, respectively. The experimental result shows that our method has a good result of 3D reconstruction for moving targets or platforms.
Laser range-gated imaging has great potentials in remote night surveillance with far detection distance and high resolution, even if under bad weather conditions such as fog, snow and rain. However, the field of view (FOV) is smaller than large objects like buildings, towers and mountains, thus only parts of targets are observed in one single frame, so that it is difficult for targets identification. Apparently, large FOV is beneficial to solve the problem, but the detection range is not available due to low illumination density in a large field of illumination matching with the FOV. Therefore, a large field-of-view range-gated laser imaging is proposed based on image fusion in this paper. Especially an image fusion algorithm has been developed for low contrast images. First of all, an infrared laser range-gated system is established to acquire gate images with small FOV for three different scenarios at night. Then the proposed image fusion algorithm is used for generating panoramas for the three groups of images respectively. Compared with raw images directly obtained by the imaging system, the fused images have a larger FOV with more detail target information. The experimental results demonstrate that the proposed image fusion algorithm is effective to expand the FOV of range-gated imaging.
Underwater 3D range-gated imaging can extend the detection range over underwater stereo cameras, and also has great potentials in real-time high-resolution imaging than 3D laser scanning. In this paper, a triangular-range-intensity profile spatial correlation method is used for underwater 3D range-gated imaging. Different from the traditional trapezoidal method, in our method gate images have triangular range-intensity profiles. Furthermore, inter-frame correlation is used for video-rate 3D imaging. In addition, multi-pulse time delay integration is introduced to shape range-intensify profiles and realize flexible 3D SRGI. Finally, in experiments, 3D images of fish net, seaweed and balls are obtained with mm-scaled spatial and range resolution.
Underwater range-gated laser imaging (URGLI) still has some problems like un-uniform light, low brightness and contrast. To solve the problems, a variant of adaptive histogram equalization called contrast limited adaptive histogram equalization (CLAHE) is proposed in this paper. In experiment, using the CLAHE and HE to enhance the images, and evaluate the quality of enhanced images by peak signal to noise ratio (PSNR) and contrast. The result shows that the HE gets the images over-enhanced, while the CLAHE has a good enhancement with compressing the over-enhancement and the influence of un-uniform light. The experimental results demonstrate that the CLAHE has a good result of image enhancement for target detection by underwater range-gated laser imaging system.
Three-dimensional super-resolution range-gated imaging (3D SRGI) is a new technique for high-resolution 3D sensing. Up to now, 3D SRGI has been developed with two range-intensity correlation algorithms, including trapezoidal algorithm and triangular algorithm. To obtain high depth-to-resolution ratio of 3D imaging, coding method was developed for 3D SRGI based on the trapezoidal algorithm in 2011. In this paper, we propose the range-intensity coding based on the triangular algorithm and the hybrid range-intensity coding based on the triangular and trapezoidal algorithms. The theoretical models to predict the maximum coding bin number are developed for different coding methods. In the models, the maximum coding bin number is 7 for three coding gate images under the triangular algorithm, and the maximum is extended to 16 under the hybrid algorithm. The coding examples of 7 bins and 16 bins mentioned above are also given in this paper. The comparison among the three coding methods is performed by the depth-to-resolution ratio defined as the ratio between the 3D imaging depth and the product of the range resolution and raw gate image number, and the hybrid coding method has the highest depth-to-resolution ratio. Higher depth-to-resolution ratio means better 3D imaging capability of 3D SRGI.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.