Handheld Spectrometer and Neural Network Model for Chemical and Biological Point Detection

A method of detecting contaminants in a chemical sample, the method comprising:

• sending, from a chemical analysis application (“app”) installed on a mobile computing device, a control signal to a handheld spectrometer apparatus connected to the mobile computing device that causes the spectrometer apparatus to perform a first test cycle involving a first sample;
• receiving, at the app and from the spectrometer apparatus, first image data captured by a photoreceptor of the spectrometer apparatus, the first image data including spectral data for the first sample;
• passing the first image data to a deep neural network (DNN) model that is trained to detect and quantify, in spectral data, one or more contaminants of a plurality of potential contaminants that include Polyalphaolefin (PAO), sulfur compound(s), synthetic fuel additive(s), hydraulic fluid(s), and microbial compound(s);
• determining, via the DNN model and based on the first image data, the first sample includes a first contaminant;
• and
• presenting, via a graphical user interface (GUI) for the app, a notification indicating the first sample includes the first contaminant.

A system for detecting contaminants in a chemical sample, the system comprising a processor and machine-readable media including instructions which, when executed by the processor, cause the processor to:

• send, from a chemical analysis application (“app”) installed on a mobile computing device, a control signal to a handheld spectrometer apparatus connected to the mobile computing device that causes the spectrometer apparatus to perform a first test cycle involving a first sample;
• receive, at the app and from the spectrometer apparatus, first image data captured by a photoreceptor of the spectrometer apparatus, the first image data including spectral data for the first sample;
• pass the first image data to a deep neural network (DNN) model that is trained to detect and quantify, in spectral data, one or more contaminants of a plurality of potential contaminants that include Polyalphaolefin (PAO), sulfur compound(s), synthetic fuel additive(s), hydraulic fluid(s), and microbial compound(s);
• determine, via the DNN model and based on the first image data, the first sample includes a first contaminant;
• and
• present, via a graphical user interface (GUI) for the app, a notification indicating the first sample includes the first contaminant.