Insect Hunter Orchard Robot Bitnap Analysis Edge Detection

Robot Features

Letting loose thousands of small predators into an environment is no different then how the cane toad problem of Australia started. Plus, you're wiping out the genetic diversity of the naturally occurring members of the same species. Genetically engineering crops to make them more resistant to toxins does not stop the dumping of toxins into the environment. It just allows them to use more herbicides to make sure they kill all the other plants. The problem with that is these toxins often mimic hormones that can affect us and especially children. We need to have someone physically go in and pull the weeds and squash the harmful insects, while leaving the good ones alone. Robots are the only solution. I may be doing Computer Engineering but my fist love is biology. I really wanted to be a biologist but I realized long ago though that there are not enough engineers who are environmentally sensitive enough.

Overall Characteristcs: Travels at a speed of  one km per hour. One robot can inspect   three acres per day. Inspects 100 acres of  crops per month. Solar-powered. Re-uses old computer processors. One centralized computer will control several smaller "dumb" robots. Frames will be processed at a rate of two frames per second.

The Set Up: A centralized computer will control several smaller dumb robots. (Unless it is found that it is more energy efficient to make them more independent, but still reasonable in price. More complex processors are generally more thermally unstable. So, on a hot summer day you could have problems.) This will allow the actual robots to be more energy efficient and cheaper to build. Vision recognition will be used to identify the "good" insect from the "bad" insects. Allowing the good ones to remain around and take care of what they are best at. I.E. the bees pollinating, the worms recycling organic material and aerating the soil, the lady beetles eating some of the not so good insects, etc... Solar energy will be used for both recharging batteries and as a supplemental power supply. Will also design the central computer station to be run off of solar power. Looking for a battery that is ideal for the highest recycle ability. I have researched the feasibility of using a communication system at 2.4 Gigahertz for the carrier wave. It appears to be a feasible method that might allow a central computer to do the image processing. Initial calculations put the power requirement to be a minimum of 7.3 milliwatts. My calculations will be shown here at a later date. Naturally, we will probably use a minimum of 1 to 5 watts to ensure a higher quality for the data transmission due to noise and interference. Testing needs to be done; however at least we know from a power point of view that the robot could transmit the images quickly enough for processing. The images will not be in a  video format, but rather a series of stills. Video is typically around 30 frames per second and I'm looking at maybe using 2 frames per second. Facts & Figures: I have come up with some numbers.  They turn out to be good, so far. It looks like one robot can inspect just over three acres per day during the summer. This requires the robots to move at only one kilometer per hour or 62% of one mile per hour. Which means that over the course of a month, a robot can inspect 100 acres of crop. (Based on the robot having a field of view of one square metre or one square yard.) According to a recent show on Genetically Modified Organisms they stated that for corn that it was generally sprayed with pesticides 7 to 8 times per season. So, if we say that this is once a month then a single robot can cover 100 acres. This does not take in to account how long the poisons last. But if 12 robots were used for every 100 acres then each plant could be visually inspected every three days. Traditional methods pretty much have to wait for a certain pest population to come into existence before it is feasible. But this means crop damage has already occurred. By have each plant checked every three days you can minimize crop damage immediately and not have to wait for a certain level of infestation. The robots would start protecting the crops from the very first day they are planted.

The most likely design the first robot will take on will be a wheeled robot for an orchard because moving between trees is much easier than moving between small plants. Realistically, more time and resources could then be spent on the image recognition system and we can explore the use of springs and other energy conservation mechanisms. The second stage will be some type of insect like robot.  Either a six or eight legged walking form as this is a very simple and stable type of body. It allows three or four legs to be on the ground while the others are being positioned. Thus, with three or four legs will always be on the ground, creating a stable base to support the robot. Two arms are likely with one arm with a camera and another with either a grabbing tool and/or a vacuum tube. However, a single arm could do just as well as the camera could easily be mounted on the collecting arm and help guide it to the insect.  As the technology progresses, biped robots might be developed, but they require more computer processing power.  Looking at the possibility of building some of the body structure out of recycled plastics and metal. Wheeled varieties are more suitable for orchards and not as likely for other crops as they make too much contact with the ground. 

The first crops will probably be either potatoes, corn, apples or some similar large leafed plant that the first prototypes won't have much of a problem walking around and moving the leaves. Tobacco may be a good crop to experiment on since the main problem I've heard is aphids, and aphids would be a great first target for the robots. Small and easily vacuumed. Also need to take into consideration the types of pests that are involved. The first robots will be slow due to energy considerations.