Identifying and monitoring productivity, health, and safety risks in industrial sites

A non-transitory computer-readable medium, wherein the non-transitory computer-readable medium is provisioned with program instructions that, when executed by at least one processor, cause a computing platform to:

• receive input data captured by one or more data input devices installed at a construction site;
• determine, from a plurality of data models, a given data model that has been trained on training data for construction sites to detect activities and objects present at the construction site;
• provide the input data to the given data model for evaluation and thereby produce output data indicating at least one given activity or object at the construction site;
• based on the output data, determine, from a set of predefined risk-prone characteristics that may be present for at least one of activities or objects at construction sites, that a first predefined risk-prone characteristic and a second predefined risk-prone characteristic of the set of predefined risk-prone characteristics are present for the at least one given activity or object at the construction site;
• based on determining that the first predefined risk-prone characteristic and the second predefined risk-prone characteristic are present for the at least one given activity or object at the construction site, determine, from a set of predefined risks, that at least one productivity, health, or safety risk of the set of predefined risks corresponds to the at least one given activity or object at the construction site;
• and
• based on determining that the at least one productivity, health, or safety risk corresponds to the at least one given activity or object at the construction site, generate one or more indications of the at least one productivity, health, or safety risk for transmission to one or more computing devices.

A method carried out by a computing platform, the method comprising:

• receiving input data captured by one or more data input devices installed at a construction site;
• determining, from a plurality of data models, a given data model that has been trained on training data for construction sites to detect activities and objects present at the construction site;
• providing the input data to the given data model for evaluation and thereby producing output data indicating at least one given activity or object at the construction site;
• based on the output data, determining, from a set of predefined risk-prone characteristics that may be present for at least one of activities or objects at construction sites, that a first predefined risk-prone characteristic and a second predefined risk-prone characteristic of the set of predefined risk-prone characteristics are present for the at least one given activity or object at the construction site;
• based on determining that the first predefined risk-prone characteristic and the second predefined risk-prone characteristic are present for the at least one given activity or object at the construction site, determining, from a set of predefined risks, that at least one productivity, health, or safety risk of the set of predefined risks corresponds to the at least one given activity or object at the construction site;
• and
• based on determining that the at least one productivity, health, or safety risk corresponds to the at least one given activity or object at the construction site, generating one or more indications of the at least one productivity, health, or safety risk for transmission to one or more computing devices.