Depth Axis Research Articles

A multiple fascicle muscle force model of the human triceps surae

Muscle is typically modelled using a lump sum idealization, scaling a single fascicle to represent the entire muscle. However, fascicles within a muscle have unique orientations, which could result in forces exerted not only in the axis running along the tendon, but also the two perpendicular axes, describing the muscle's width and depth. The purpose of this research was to develop a geometric-based model of the soleus, medial gastrocnemius, and lateral gastrocnemius as distributed force systems which can predict three-dimensional forces. Measurements were taken from the triceps surae in two human cadavers (80 and 85 years old). These models predicted muscle volumes and ankle plantar flexor moments that were realistic considering the age of the cadavers. Small differences were observed in calcaneal tendon force and moment for the distributed force models compared to modelling muscle force using a lump sum idealization. The major finding of the distributed force models was that forces were present in the axes corresponding to the muscle's length, width, and depth. The forces in the width and depth axes may be relevant for evaluating how muscle shape changes during contraction, as well as to investigate stress-strain patterns along the muscle's proximal and distal aponeuroses.

Reverberant 3D optical coherence elastography maps the elasticity of individual corneal layers

The elasticity mapping of individual layers in the cornea using non-destructive elastography techniques advances diagnosis and monitoring of ocular diseases and treatments in ophthalmology. However, transient Lamb waves, currently used in most dynamic optical coherence and ultrasound elastography techniques, diminish the translation of wave speed into shear/Young’s modulus. Here, we present reverberant 3D optical coherence elastography (Rev3D-OCE), a novel approach leveraging the physical properties of diffuse fields in detecting elasticity gradients not only in the lateral direction, but also along the depth axis of the cornea. A Monte Carlo analysis, finite element simulations, and experiments in layered phantoms are conducted to validate the technique and to characterize the axial elastography resolution. Experiments in ex vivo porcine cornea at different intraocular pressures reveal that Rev3D-OCE enables the elastic characterization of single layers that matches the anatomical description of corneal layers with unprecedented contrast in the dynamic OCE field.

Multivariate Analysis of BOLD Activation Patterns Recovers Graded Depth Representations in Human Visual and Parietal Cortex.

Navigating through natural environments requires localizing objects along three distinct spatial axes. Information about position along the horizontal and vertical axes is available from an object’s position on the retina, while position along the depth axis must be inferred based on second-order cues such as the disparity between the images cast on the two retinae. Past work has revealed that object position in two-dimensional (2D) retinotopic space is robustly represented in visual cortex and can be robustly predicted using a multivariate encoding model, in which an explicit axis is modeled for each spatial dimension. However, no study to date has used an encoding model to estimate a representation of stimulus position in depth. Here, we recorded BOLD fMRI while human subjects viewed a stereoscopic random-dot sphere at various positions along the depth (z) and the horizontal (x) axes, and the stimuli were presented across a wider range of disparities (out to ∼40 arcmin) compared to previous neuroimaging studies. In addition to performing decoding analyses for comparison to previous work, we built encoding models for depth position and for horizontal position, allowing us to directly compare encoding between these dimensions. Our results validate this method of recovering depth representations from retinotopic cortex. Furthermore, we find convergent evidence that depth is encoded most strongly in dorsal area V3A.

Superlocalized Three-Dimensional Live Imaging of Mitochondrial Dynamics in Neurons Using Plasmonic Nanohole Arrays.

We investigated the transport of neuronal mitochondria using superlocalized near-fields with plasmonic nanohole arrays (PNAs). Compared to traditional imaging techniques, PNAs create a massive array of superlocalized light beams and allow 3D mitochondrial dynamics to be sampled and extracted almost in real time. In this work, mitochondrial fluorescence excited by the PNAs was captured by an optical microscope using dual objective lenses, which produced superlocalized dynamics while minimizing light scattering by the plasmonic substrate. It was found that mitochondria move with an average velocity 0.33 ± 0.26 μm/s, a significant part of which, by almost 50%, was contributed by the movement along the depth axis ( z-axis). Mitochondrial positions were acquired with superlocalized precision (σ x = 5.7 nm and σ y = 11.8 nm) in the lateral plane and σ z = 78.7 nm in the z-axis, which presents an enhancement by 12.7-fold in resolution compared to confocal fluorescence microscopy. The approach is expected to serve as a way to provide 3D information on molecular dynamics in real time.

Quantification of Cretaceous-Cenozoic exhumation in the Otway Basin using sonic velocities and implications for hydrocarbon exploration

Increasing demand for hydrocarbons is prompting increasing exploration in poorly understood and geologically complex basins. These include basins affected by exhumation, where there is often uncertainty regarding the maximum burial depths of source, reservoir and seal horizons. This uncertainty can be reduced by quantifying exhumation magnitudes using techniques such as sonic velocity analysis. The Cretaceous- Cenozoic Otway Basin of the southern Australian margin is an important oil and gas province whose subsidence history was interrupted by exhumation events during the mid-Cretaceous, mid-Eocene and late Miocene. We have used sonic velocity data from 135 wells to quantify the magnitude and extent of exhumation during these events. Here we present estimates of net exhumation based on analyses of Lower Cretaceous Eumeralla Formation shales, which are important source rocks within the basin. Our approach involves measuring the velocities of overcompacted intra-formational shale units identified from gamma ray logs, and determining their displacement on the depth axis from an empirically derived normal compaction trend based on data from wells where the Eumeralla Formation is at maximum burial at presentday. Our results indicate significant values of mid- Cretaceous net exhumation in the Otway Ranges and Colac Trough, with estimates for Olangolah-1 (~2700 m), Stonyford-1 (~1100 m), Tirrengowa-1 (~860 m) and Ingleby-1 (~1000 m) in good agreement with independent estimates of mid-Cretaceous exhumation for these wells based on palaeothermal data. There is also moderate net exhumation around the Merino High (~700 m) but minor (0-200m) net exhumation around the basin?s northern margins and Penola Trough area and moderate (200-600 m) around the Port Campbell Embayment gas fields. Mid-Cretaceous exhumation is interpreted to have been caused by NW-SE directed compression and inversion of the basin. Our results indicate that potential Eumeralla Formation source rocks in wells where large values of mid-Cretaceous exhumation are recorded are most likely over-mature for hydrocarbon generation, in agreement with previous studies of the thermal history of the basin.

Rearview Camera-Based Stixel Generation for Backing Crash Prevention

Backing crashes have become a serious problem, resulting in the United States announcing a rule requiring all vehicles to come equipped with rearview cameras. As a result of this situation, this paper proposes a method that detects obstacles behind a vehicle using a rearview camera with a fisheye lens. The proposed method represents obstacles by means of the Stixel World, and it consists of two stages: free space boundary detection and obstacle height estimation. In the first stage, obstacles are three-dimensionally reconstructed, and the 3D points are spatially interpolated and sequentially accumulated. An occupancy grid map is generated with depth and incidence angle axes to effectively handle a wide field-of-view situation. In the second stage, this method combines some complementary information to reliably estimate the obstacle height using spare 3D points. Three cost maps based on 3D points, color similarity, and edge strength are calculated and fused in height and incidence angle axes. The free space boundary and the obstacle height are estimated by finding the optimal paths in the occupancy grid map and the fused cost map, respectively. Finally, Stixels are generated by integrating the free space boundary and obstacle height. In the experiment, the proposed method was quantitatively evaluated and compared with the previous methods in various road, obstacle, and background conditions.

Scaling laws of convection for cooling planets in a stagnant lid regime

Scaling laws used in 1-D parametrized thermal models to describe the cooling of planets have mainly been derived from steady-state convection models. Here, fully dynamic 2-D and 3-D models are employed to compute the interior thermal evolution using strong temperature-dependent viscosity, mixed heating with radioelement decay, spherical geometry. The results of these models are compared to those of parametrized models, where we distinguish between the upper thermal boundary layer (TBL) and the stagnant lid, in a Monte Carlo framework to derive scaling laws valid for the entire planet evolution.The upper TBL thickness depends on the Rayleigh number to the power βu, while the lid base is determined by the mantle temperature and the rate of viscosity change with temperature multiplied by a prefactor arh. By varying βu and arh we find that, although the heating conditions change as a function of time, the thermal evolution of a cooling planet in a stagnant lid regime can be represented by a wide range of combinations of arh and βu. Suitable fits are found for different values of arh and βu that vary depending on the model set-up. The observed relationship between suitable values of arh and βu is explained by a trade-off between the lid and the TBL thicknesses. Only when considering a specific definition of the stagnant lid thickness in 3-D simulations, suitable combinations for all models converge to specific values of arh and βu. Using a stagnant lid defined by the intersection between the depth axis and the tangent to the velocity profile at the depth corresponding to the maximum velocity gradient, we find best-fit values of 2.54 and 0.335 for arh and βu, respectively.

Relationship between testicular and epididymal biometry compared to semen quality in Yankasa rams

This study examined the relationship between the testicular and epididymal biometry compared to semen characteristics in Yankasa rams. Fifteen apparently healthy rams aged between 1-2 years with average weight of 20 kg were used for the study. They were acclimated for four weeks, screened and treated against endo and ecto parasite. They were kept in small ruminant pens under zero grazing, fed with hay, ground nut leaves/straw ‘harawa’ and wheat offal with water supplied ad-libitum. The scrotal circumference was measured in centimeters using a measuring tape. The testicular length (L), width (W), depth and epididymal head axis, length and tail axis were measured by caliper and testicular volume and weight were calculated by formula methods. Semen samples were collected using an electro-ejaculator and immediately evaluated for colour, volume, motility, pH, sperm concentration, sperm morphology and live/dead ratio. The right testicular and epidydimal biometry were non significantly greater than that of the left. Semen parameters correlate positively (right testicules, r=0.225) and (left testicles, r=0.346) with testicular volume except for bent tail, coil tail and motility. The left epididymal head’s long axis and right epididymal tail axis showed significant positive correlation (r= 0.55) with semen parameters and right epididymal head’s long axis and left epididymal tail axis showed negative correlation with semen characteristics. It was concluded that testicular and epididymal biometry increase with increase testicular function and output.

Multiple distance cues do not prevent systematic biases in reach to grasp movements.

The perceived distance of objects is biased depending on the distance from the observer at which objects are presented, such that the egocentric distance tends to be overestimated for closer objects, but underestimated for objects further away. This leads to the perceived depth of an object (i.e., the perceived distance from the front to the back of the object) also being biased, decreasing with object distance. Several studies have found the same pattern of biases in grasping tasks. However, in most of those studies, object distance and depth were solely specified by ocular vergence and binocular disparities. Here we asked whether grasping objects viewed from above would eliminate distance-dependent depth biases, since this vantage point introduces additional information about the object's distance, given by the vertical gaze angle, and its depth, given by contour information. Participants grasped objects presented at different distances (1) at eye-height and (2) 130mm below eye-height, along their depth axes. In both cases, grip aperture was systematically biased by the object distance along most of the trajectory. The same bias was found whether the objects were seen in isolation or above a ground plane to provide additional depth cues. In two additional experiments, we verified that a consistent bias occurs in a perceptual task. These findings suggest that grasping actions are not immune to biases typically found in perceptual tasks, even when additional cues are available. However, online visual control can counteract these biases when direct vision of both digits and final contact points is available.

Deep-sea mobile megafauna of Mediterranean submarine canyons and open slopes: Analysis of spatial and bathymetric gradients

Mediterranean canyons are generally characterized by complex topography and hydrographic conditions, and can act as conduits for organic matter and sediments from the shelf to the deep sea, which result in enhanced macro- and megafaunal abundance and biodiversity. Here, we reviewed the literature of the last 20 years and combined available information with new data, from our own datasets (including unpublished data), on megafaunal species inhabiting submarine canyons and the adjacent slopes. We analysed the data acquired using different gears, from baited cameras on aluminum frames and landers and baited traps to commercial trawl gears, in order to combine and complement the results in megafaunal biodiversity and assemblage structure in Mediterranean canyons. Overall, the results of our analysis on megafaunal assemblages allowed identifying 400 species, including 28 species of cephalopods, 64 of decapod crustaceans, 25 of echinoderms, 27 of elasmobranches and 127 of Actinopterygii. We observed significant longitudinal differences in assemblage composition, with high species turnover across Mediterranean sub-basins. Our analysis also confirmed the importance of depth axis as the main factor influencing the structure of the deep-sea demersal assemblages, with greater diversity in the upper part of the slopes. The multivariate multiple regression analysis showed that geomorphological features of the canyons (i.e., slope, dendricity, sinuosity, and distance from adjacent canyons) contribute significantly to explain the observed variability in the taxonomic composition of megafaunal assemblages at sub-basin scale. The results presented here are important in global biodiversity management perspectives, and confirm that canyons represent a seascape unit deserving protection at different levels modulated according to biogeographic and geomorphological features.

Comparisons of Target Localization Abilities during Physical and Virtual Rotating Scenes by Cognitively-Intact and Cognitively Impaired Older Adults

Background: Previous studies have reported that coordinate information (i.e. distance between any two objects in a specific direction) is encoded differently from Virtual Reality (VR) and physical scenes. However, the accuracy of encoding categorical information (i.e. relative positions of objects) from VR scenes has not been adequately investigated. In this study, we used a novel rotating visual scene to study the effects of aging, prior experience with VR, and dementia on the accuracy of encoding categorical information between physical and virtual environments. Methods: We recruited a cohort of 60 cognitively-healthy older adults with and without previous VR experience (Experiment 1) as well as 18 older adults with mild to moderate Alzheimer disease (AD) (Experiment 2). In both of the experiments, the participants were asked to attend to a target window in a virtual or real building (dependent to the group assignment) as the building is being rotated around its vertical axis in depth of the scene. They were required to verbally judge the final position of the target in terms of directions (e.g., left, right, back, and front) with respect to the entrance of the buildings after the full rotation has stopped. The experimenters calculated a score for each participants based on s/her accuracy in locating the target window. Results: Healthy older adults succeeded in accurately localizing the target's position from both environments, whereas individuals with AD were only able to encode the target’s position from the physical environment. Conclusions: Our results suggest the inability to encode from a rotating VR scene might be a symptom of dementia

How Deep Is Your SNARC? Interactions Between Numerical Magnitude, Response Hands, and Reachability in Peripersonal Space.

Spatial, physical, and semantic magnitude dimensions can influence action decisions in human cognitive processing and interact with each other. For example, in the spatial-numerical associations of response code (SNARC) effect, semantic numerical magnitude facilitates left-hand or right-hand responding dependent on the small or large magnitude of number symbols. SNARC-like interactions of numerical magnitudes with the radial spatial dimension (depth) were postulated from early on. Usually, the SNARC effect in any direction is investigated using fronto-parallel computer monitors for presentation of stimuli. In such 2D setups, however, the metaphorical and literal interpretation of the radial depth axis with seemingly close/far stimuli or responses are not distinct. Hence, it is difficult to draw clear conclusions with respect to the contribution of different spatial mappings to the SNARC effect. In order to disentangle the different mappings in a natural way, we studied parametrical interactions between semantic numerical magnitude, horizontal directional responses, and perceptual distance by means of stereoscopic depth in an immersive virtual reality (VR). Two VR experiments show horizontal SNARC effects across all spatial displacements in traditional latency measures and kinematic response parameters. No indications of a SNARC effect along the depth axis, as it would be predicted by a direct mapping account, were observed, but the results show a non-linear relationship between horizontal SNARC slopes and physical distance. Steepest SNARC slopes were observed for digits presented close to the hands. We conclude that spatial-numerical processing is susceptible to effector-based processes but relatively resilient to task-irrelevant variations of radial-spatial magnitudes.

From top to bottom: Do Lake Trout diversify along a depth gradient in Great Bear Lake, NT, Canada?

Depth is usually considered the main driver of Lake Trout intraspecific diversity across lakes in North America. Given that Great Bear Lake is one of the largest and deepest freshwater systems in North America, we predicted that Lake Trout intraspecific diversity to be organized along a depth axis within this system. Thus, we investigated whether a deep-water morph of Lake Trout co-existed with four shallow-water morphs previously described in Great Bear Lake. Morphology, neutral genetic variation, isotopic niches, and life-history traits of Lake Trout across depths (0–150 m) were compared among morphs. Due to the propensity of Lake Trout with high levels of morphological diversity to occupy multiple habitat niches, a novel multivariate grouping method using a suite of composite variables was applied in addition to two other commonly used grouping methods to classify individuals. Depth alone did not explain Lake Trout diversity in Great Bear Lake; a distinct fifth deep-water morph was not found. Rather, Lake Trout diversity followed an ecological continuum, with some evidence for adaptation to local conditions in deep-water habitat. Overall, trout caught from deep-water showed low levels of genetic and phenotypic differentiation from shallow-water trout, and displayed higher lipid content (C:N ratio) and occupied a higher trophic level that suggested an potential increase of piscivory (including cannibalism) than the previously described four morphs. Why phenotypic divergence between shallow- and deep-water Lake Trout was low is unknown, especially when the potential for phenotypic variation should be high in deep and large Great Bear Lake. Given that variation in complexity of freshwater environments has dramatic consequences for divergence, variation in the complexity in Great Bear Lake (i.e., shallow being more complex than deep), may explain the observed dichotomy in the expression of intraspecific phenotypic diversity between shallow- vs. deep-water habitats. The ambiguity surrounding mechanisms driving divergence of Lake Trout in Great Bear Lake should be seen as reflective of the highly variable nature of ecological opportunity and divergent natural selection itself.

Active Depth Cuts Without Distortion of Stereoscopic Depth Reduce Annoyance

In stereo 3D (S3D) production, an active depth cut means to shift the presented scene along the depth axis in order to mitigate strong depth discontinuities at scene cuts, which would otherwise induce visual fatigue to the human visual system. This is done using the dynamic horizontal image translation (DHIT) approach, where the S3D views are slowly shifted in opposite directions to alter the distance to the perceived scenery. A problem of the DHIT is that the presented depth is distorted during the process. In this paper, a distortion-free DHIT approach is proposed. The distortions are eliminated by dynamically adjusting the distance between the two stereo views. The new approach is compared with the regular DHIT in a subjective experiment with the result that it is significantly less perceivable and annoying. The critical speeds for annoyance and perception with both approaches are reported, and an active depth cut design recommendation for S3D producers is given.

3D topology of orientation columns in visual cortex revealed by functional optical coherence tomography.

Orientation tuning is a canonical neuronal response property of six-layer visual cortex that is encoded in pinwheel structures with center orientation singularities. Optical imaging of intrinsic signals enables us to map these surface two-dimensional (2D) structures, whereas lack of appropriate techniques has not allowed us to visualize depth structures of orientation coding. In the present study, we performed functional optical coherence tomography (fOCT), a technique capable of acquiring a 3D map of the intrinsic signals, to study the topology of orientation coding inside the cat visual cortex. With this technique, for the first time, we visualized columnar assemblies in orientation coding that had been predicted from electrophysiological recordings. In addition, we found that the columnar structures were largely distorted around pinwheel centers: center singularities were not rigid straight lines running perpendicularly to the cortical surface but formed twisted string-like structures inside the cortex that turned and extended horizontally through the cortex. Looping singularities were observed with their respective termini accessing the same cortical surface via clockwise and counterclockwise orientation pinwheels. These results suggest that a 3D topology of orientation coding cannot be fully anticipated from 2D surface measurements. Moreover, the findings demonstrate the utility of fOCT as an in vivo mesoscale imaging method for mapping functional response properties of cortex in the depth axis. NEW & NOTEWORTHY We used functional optical coherence tomography (fOCT) to visualize three-dimensional structure of the orientation columns with millimeter range and micrometer spatial resolution. We validated vertically elongated columnar structure in iso-orientation domains. The columnar structure was distorted around pinwheel centers. An orientation singularity formed a string with tortuous trajectories inside the cortex and connected clockwise and counterclockwise pinwheel centers in the surface orientation map. The results were confirmed by comparisons with conventional optical imaging and electrophysiological recordings.

3D forward modeling of upgoing and downgoing wavefields using Hilbert transform

The separation of upgoing and downgoing wavefields is an important technique in the processing of vertical seismic profiling data and ocean bottom cable data. It is also used in reverse time migration (RTM) based on the two-way wave equation to suppress low-frequency, high-amplitude noises and false images. Therefore, we model upgoing and downgoing wavefields directly in the wavefield propagation. There are several methods to obtain separated wavefields. The methods using the Fourier transform require storage of the wavefields, which is not practical due to the extremely high disk-space requirements. Methods using Poynting vectors have an ambiguity problem when crossing a peak or a trough of the wavefields. To improve the accuracy and stability of the modeled upgoing and downgoing wavefields in a complicated velocity model, we evaluate an efficient forward-modeling approach purely based on the Hilbert transform in 3D acoustic wavefield simulation. This method is implemented by the Hilbert transform along the time and depth axis, instead of the Fourier transform. We explicitly derive the formulas for upgoing and downgoing wavefield propagation and attach reproducible source codes. Applications to synthetic models indicate that this method can forward propagate upgoing and downgoing wavefields effectively and improve the imaging quality in migration. This method has various potential applications, e.g., 3D seismic imaging with high computation efficiency.

Shear excitation of a multilayered magneto-electro-elastic half-space considering a vast frequency content

The objective of this paper is to present dynamic responses of a multilayered magneto-electro-elastic half-space due to an external/internal time-harmonic shear mechanical loading. To do so, a concrete mathematical tool is applied to solve a set of complicated high-order coupled equations of motion for a layered magneto-electro-elastic structure. The mathematical tool is constructed based on a new infinite dimensional vector space utilizing multiplication of exponential functions as a base for Fourier series in the tangential direction and Bessel function as a base in the radial direction. A cylindrical coordinate system is attached to the multi-layered half-space with the depth axis oriented perpendicular to the surface. In this way, with the use of this vector space, the coupled governing partial differential equations are changed to two sets of ordinary first-order differential equations, where their unknown functions are the coefficients of base functions in the Hankel-Fourier domain. It should be noted that these two sets of equations are completely decoupled based on the decoupling of SH-wave from P- and SV-waves. The effect of the layering is considered with the use of the dual-variable-position method. A mechanically uniform surface/internal shear time-harmonic loading is applied as a common excitation in nature to pursue the pattern of energy transformation in multifunctional magneto-electro-elastic structures with different material properties. Also, the effects of loading frequency and resonance phenomena are investigated in this paper numerically. The results show clearly how the arrangement and the designing process of different layers with different material properties can change responses in smart devices.

Benthic litter distribution on circalittoral and deep sea bottoms of the southern Bay of Biscay: Analysis of potential drivers

We analyze marine litter densities in soft bottoms of the southern Bay of Biscay using five years of demersal trawling data (2006–2010). Marine litter densities amounted to 43 ± 33kgkm−2 and 74 ± 28 itemskm−2, with plastics and fisheries derived litter being the most widespread categories. Litter densities generally decreased along the water depth axis. To identify possible drivers for the observed litter distribution we performed a generalised additive model, which explained 14.8% of the variance and pointed to densely populated areas, number of fishing ports, geographical sector and fishing activity as the main explanatory factors. The most important driver for the benthic litter distribution was human population, as litter density linearly increased along this variable. Similarly, the number of ports in neighbouring areas had a positive effect on litter densities. Fishing effort had a negative and non-linear effect on benthic litter density which could be explained by litter delocalization during fishing operations. We hypothesise that litter might accumulate preferentially on the periphery of rocky bottoms, out of reach for our sampling methodology. Litter distribution differed among geographical sectors, pointing to other variables such as shipping traffic and oceanographic currents, which were not explicitly considered in the analysis. Our study sets a reference level for benthic macro-litter in the southern Bay of Biscay and identifies factors driving its distribution, which can be extrapolated to other continental shelf seas. Our findings lay the foundations to develop measures aiming to reduce macro-litter densities on the seafloor.

A study on validating KinectV2 in comparison of Vicon system as a motion capture system for using in Health Engineering in industry

AbstractThe currently available commercial motion capture systems are constrained by space requirement and thus pose difficulties when used in developing kinematic description of human movements within the existing manufacturing and production cells. The Kinect sensor does not share similar limitations but it is not as accurate. The proposition made in this article is to adopt the Kinect sensor in to facilitate implementation of Health Engineering concepts to industrial environments. This article is an evaluation of the Kinect sensor accuracy when providing three dimensional kinematic data. The sensor is thus utilized to assist in modeling and simulation of worker performance within an industrial cell. For this purpose, Kinect 3D data was compared to that of Vicon motion capture system in a gait analysis laboratory. Results indicated that the Kinect sensor exhibited a coefficient of determination of 0.9996 on the depth axis and 0.9849 along the horizontal axis and 0.2767 on vertical axis. The results prove the competency of the Kinect sensor to be used in the industrial environments.

An automatic recognition and parameter extraction method for structural planes in borehole image

As a breakthrough in borehole imaging technology, digital panoramic borehole camera technology has been widely employed. The high-resolution panoramic borehole images can accurately reproduce the geometric features of structural planes. However, the detection of these features is usually done manually, which is both time-consuming and introduces human errors. To solve this problem, this paper presents a method for the automatic recognition and parameter extraction of borehole geometric features of camera images. In this method, the image's gray and gradient level, and also their projection on the depth axis are used to identify the locations of structural planes. Afterwards, iterative matching is employed by using a template of sinusoidal function to search for structural planes in the identified image blocks. Finally, optimal sine curves are selected as the feature curves of structural planes, and their related parameters are converted into structural plane parameters required for engineering, such as their positions, dip directions, dip angles and fracture widths. The method can automatically identify all of structural planes throughout the whole borehole camera image in a continuous and rapid manner, and obtain the corresponding structural parameters. It has proven highly reliable, accurate and efficient.