Label-free bio-aerosol sensing using mobile microscopy and deep learning
A label-free bio-aerosol sensing platform and method uses a field-portable and cost-effective device based on holographic microscopy and deep-learning, which screens bio-aerosols at a high throughput level. Two different deep neural networks are utilized to rapidly reconstruct the amplitude and phase images of the captured bio-aerosols, and to output particle information for each bio-aerosol that is imaged. This includes, a classification of the type or species of the particle, particle size, particle shape, particle thickness, or spatial feature(s) of the particle. The platform was validated using the label-free sensing of common bio-aerosol types, e.g., Bermuda grass pollen, oak tree pollen, ragweed pollen, Aspergillus spore, and Alternaria spore and achieved >94% classification accuracy. The label-free bio-aerosol platform, with its mobility and cost-effectiveness, will find several applications in indoor and outdoor air quality monitoring.
1. A method of classifying aerosol particles using a portable microscope device comprising:
capturing aerosol particles on an optically transparent substrate;
illuminating the optically transparent substrate containing the captured aerosol particles with one or more illumination sources contained in the portable microscope device;
capturing holographic images or diffraction patterns of the captured aerosol particles with an image sensor disposed in the portable microscope device and disposed adjacent to the optically transparent substrate;
processing the image files containing the holographic images or diffraction patterns with image processing software contained on a local or remote computing device, wherein image processing comprises inputting the holographic images or diffraction patterns through a first trained deep neural network of the image processing software to output reconstructed amplitude and phase images of each aerosol particle at the one or more illumination wavelengths and wherein a second trained deep neural network of the image processing software receives as an input the outputted reconstructed amplitude and phase images of each aerosol particle at the one or more illumination wavelengths and outputs one or more of the following for each aerosol particle: a classification or label of the type of aerosol particle, a classification or label of the species of the aerosol particle, a size of the aerosol particle, a shape of the aerosol particle, a thickness of the aerosol particle, and a spatial feature of the particle.
2. The method of claim 1 , wherein capturing aerosol particles comprises activating a vacuum pump disposed in the portable microscope device.
3. The method of claim 1 , wherein the image files containing the holographic images or diffraction patterns are transferred from the portable microscope device to a remote computing device containing the image processing software.
4. The method of claim 1 , wherein the image files containing the holographic images or diffraction patterns are processed using a computing device that is integrated within the portable microscope device.
5. The method of claim 1 , wherein the image files containing the holographic images or diffraction patterns are processed using a computing device that is locally connected to the portable microscope device.
6. The method of claim 5 , wherein the computing device is locally connected to the portable microscope device via a wireless or wired connection.
7. The method of claim 1 , wherein the aerosol particles comprise bio-aerosol particles.
8. The method of claim 1 , wherein the one or more illumination sources comprises a one or more laser diodes and/or one or more light emitting diodes (LEDs).
9. The method of claim 1 , wherein the input to the second trained deep neural network comprises cropped images of the reconstructed amplitude and phase images of each aerosol particle at the one or more illumination wavelengths.
10. A system for classifying aerosol particles comprising:
a portable, lens-free microscopy device for monitoring air quality comprising:
a housing;
a vacuum pump configured to draw air into an impaction nozzle disposed in the housing, the impaction nozzle having an output located adjacent to an optically transparent substrate for collecting particles contained in the air;
one or more illumination sources disposed in the housing and configured to illuminate the collected particles on the optically transparent substrate;
an image sensor disposed in the housing and located adjacent to the optically transparent substrate, wherein the image sensor collects diffraction patterns or holographic images cast upon the image sensor by the collected particles;
a computing device comprising one or more processors executing image processing software thereon and configured to receive the holographic images or diffraction patterns obtained from the portable, lens-free microscopy device, wherein the image processing software inputs the holographic images or diffraction patterns obtained at the one or more illumination wavelengths through a first trained deep neural network to output reconstructed amplitude and phase images of each aerosol particle and inputs the reconstructed amplitude and phase images of each aerosol particle in a second trained deep neural network and outputs one or more of the following for each aerosol particle: a classification or label of the type of aerosol particle, a classification or label of the species of the aerosol particle, a size of the aerosol particle, a shape of the aerosol particle, a thickness of the aerosol particle, and a spatial feature of the particle.
11. The system of claim 10 , wherein the computing device is located remote from the portable, lens-free microscopy device.
12. The system of claim 10 wherein the computing device is locally connected with the portable, lens-free microscopy device via a wired or wireless communication link.
13. The system of claim 10 , wherein the computing device comprises one or more processors integrated in the portable, lens-free microscopy device.
14. The system of claim 13 , wherein the one or more processors integrated in the portable, lens-free microscopy device are configured to control the vacuum pump and the one or more illumination sources.
15. The system of claim 10 , further comprising a mobile computing device having software or an application contained thereon for controlling the operation of the portable, lens-free microscopy device and displaying images of the aerosol particles and/or classification, size, shape, thickness, or spatial feature data.