• search navSök
  • svenska
  • Logga in
Add your Description here.

    • Begär en offert

Artificial intelligence analysis & automation

Data digs deeper

The advantages of A.I. automation are speed and consistency. As miners and explorers, we are entering the age of big data, which is too voluminous, and produced too rapidly, to be interpreted by conventional methods alone. At ALS, we work together with your geoscience team to add A,I. tools and automation workflows, allowing your expert geoscientists to spend their time thinking about geoscience.

The ALS Geoanalytics difference

We believe that good data science must critically be underpinned with good geological understanding. Our team of geoscientists, with expertise in geochemistry, geophysics, and structural geology, work closely with our data science team to ensure the right inputs are being used, and the right questions are being answered.

Machine Assisted Mapping

There are many remote sensing methods that can be utilised to help the explorer understand the project area before initiating a field campaign. Prioritising different areas (target lithology, outcrop abundance) can help plan traverses and aid drill campaigns. Terrain maps can be created from clustered multispectral data. Different multispectral bands, or ratios of bands, which correspond to specific mineralogy, enables interpretation and mapping of regolith units. Outcrops can be identified by extracting textural metrics from orthophotography or LiDAR. Predictions regarding the likelihood of mineralisation (prospectivity analysis) can be made using supervised machine learning models built on a data cube of engineered feature layers.


Computer Vision

Computer vision refers to the ability of computers and artificial intelligence systems to interpret and analyse visual information from the real world, such as images and videos. It involves the use of cameras, algorithms, and software to extract information from visual data, which can be used for various applications such as quality control, autonomous navigation, and object recognition. Computer vision helps automate processes and make decisions based on visual data, improving efficiency and accuracy in various industries.


Automated Logging

Core is often relogged when a new company takes over a deposit, or when there is a lot of historical core logged by numerous geologists over the years with limited control over consistency. The core logging process is typically based on visual observations of core, which makes it an ideal candidate for automation through machine learning. Automating the logging process saves time, money and ensures an objective end result. The re-logging process requires selecting a number of training points (typically several hundred). These points, and the associated core photos, are then analysed by our proprietary algorithm to generate a data model. The model can then be applied to classify all core photos collected over time. Additional training data can be provided to improve accuracy.


Geophysical Inversion Software

Our inversion algorithm takes advantages of new developments in ML, particularly Gaussian processes for modern machine learning systems. It can invert either gravity or magnetic data. The benefit of incorporating ML into geophysical inversions is that ML algorithms can find non-linearities within the data, and therefore produce better 3d models.


Anomaly Detection and Pattern Matching

One of the strengths of A.I. is flexibility and reproducibility with multiple data types. Our pattern recognition and anomaly detection tools utilise electromagnetics, gradient magnetics, and gravity data. Anomaly matching uses profiles from known orebodies or target formations to compare with anomalies in other datasets and ranks them based on their similarity. The anomaly detection routine, or PeakFinder, is used to locate minima and maxima in line data. Peaks are discriminated based on several factors, including the peak amplitude, prominence, and width, among other attributes. The peak characteristics are used to generate several metrics such as maximum amplitude, tau decay, dip direction and angle. The anomaly matching tool examines line/channel data and compares it to responses selected from the survey data and/or those generated from theoretical/synthetic plates. Magnetic, electromagnetic and/or gravity data can be modelled. A normalised result is returned with values between 0 and 1 depending on how similar the channel data is to the target signal. The results are displayed as a similarity map that highlights segments of flight lines that are similar or dissimilar to the target signal.


 
v.1.1.1140