– John Deere, the Moline, Ill. farm-machine company, may seem like a surprising presence near the office towers of software giants like Salesforce and LinkedIn. And even though its new lab in the startup-heavy South of Market neighborhood in San Francisco is focused on technologies like artificial intelligence and automation, some devoted customers can’t stay away.

“We have had people knocking on the glass trying to buy tractors,” said Alex Purdy, head of John Deere Labs. “I actually had someone the other day who tried to come in to replace something on his mower. People don’t necessarily think about Deere as the software development shop for agriculture.”

Deere’s San Francisco outpost opened in May, and in September the company made its first big move to beef it up, agreeing to pay $305 million to buy Blue River Technology, a Sunnyvale, Calif., startup developing farm equipment using computers and robotics to automatically detect every single plant on a farm. Some of Blue River’s employees will join the San Francisco lab.

The move by 180-year-old Deere is the latest sign of agricultural giants’ focus on automation and robotics. For example, in August, DuPont bought Granular Inc., a San Francisco agriculture analytics software firm, for $300 million. Meanwhile, Deere competitor Kubota Tractor Corp. opened a new research and development facility earlier this year in its Grapevine, Texas, headquarters.

“Larger farms producing a great deal of grain or corn or other row crops are using technology with a good deal of enthusiasm,” said Will Rodger, director of policy communications for the American Farm Bureau. “Typically, the younger, better-educated farmers are more bullish on these new technologies.”

Deere has about 1,000 employees working on high-tech hardware and software worldwide. Purdy said the company hopes to use the San Francisco office, where it will have eight to 15 employees, as a “listening post for other startups” and form partnerships with other agriculture-related tech companies in the Bay Area. It intends to reach engineers versed in robotics, artificial intelligence, machine learning and cloud-based systems.

“If you’re a machine-learning engineer, you are sorting somebody’s contact list,” said Willy Pell, Blue River’s director of new technology. “You could do that, or you could come and make food growth more efficient.”

Blue River was formed seven years ago by Stanford University graduate students Jorge Heraud and Lee Redden. It has developed machinery, currently pulled by tractor through a field, that uses high-definition cameras to scan each plant, determine what kind it is and whether it is healthy. The system can also detect weeds.

There, the machinery could potentially determine whether a plant needs a spray of pesticide, herbicide or water, saving farmers from over-spraying, with its environmental and economic consequences.

Farmers such as Billy Tiller hope the deal between Blue River and John Deere will lead to technology that finds and eradicates a type of weed that can threaten cotton crops and has become resistant to chemical herbicides.

Blue River tested its equipment this summer on a portion of Tiller’s 6,300-acre farm about 60 miles west of Lubbock, Texas. Tiller said he is encouraged by even the preliminary results because he figures he could reduce the use of chemicals by more than 90 percent by using the system, which identifies, targets and delivers a herbicide hit “like a heat-seeking missile.”

“With something like a Blue River, you can get in early when the weeds are small and get those treated,” Tiller said. “We’ve got to have answers. More chemicals, even different chemicals, are not the answer to these resistant weeds.”

The company hopes to have 10 of its machines ready by next year’s cotton-planting season, Pell said.

“We have to do it really fast, and our accuracy requirements are very high,” he said. “And the things that we’re detecting are subtly different. It’d be like detecting you from your brother. It’s not like detecting a zebra from a fire truck.”

Deere is also working more broadly on automation technology, Purdy said — though fully self-driving tractors remain far in the future.

The farmer still needs to “stay in charge, ready to intervene, during critical tasks such as harvesting,” according to a report released Thursday by IDTechEx Research of Cambridge, England.

But the development of robotics that can identify the needs of individual plants could revolutionize the agricultural chemicals business, according to IDTechEx.

“Now we can see the silhouette of the agrobots of the future: small intelligent autonomous mobile robots taking precise action on an individual plant basis,” the report said. “These robots can be connected to the cloud to share learning and data, and to receive updates en mass.”

Tiller, 53, a fourth-generation cotton farmer, is enthusiastic about the prospects of more technology.

“It’s like riding a space shuttle today when you get in these new combines and these new tractors,” he said. “There’s more technology on tractors today than on that first space shuttle.’’